• image

    Digital Manufacturing Capacity Development uisng AI

    Funded by Innovated UK and AM3D

    Funded by Innovated UK for the development of a new digital manufacturing capability and embedding new technology and knowledge in additive manufacturing and 3D image processing using AI.

  • image

    Modelling and intelligent control of battery-supercapacitor hybrid energy storage systems in electric vehicles

    Funded by Royal Academy of Engineering in partnership with University of Pretoria, Average Technologies (Pty) Ltd and Green Scooter (Pty) Ltd in South Africa

    The planned research aims to provide a solution that reduces the capital cost, improves efficiency, and prolongs the lifespan of energy storage systems used in all sizes of EVs, realized in two stages. Firstly, thermal-electric modelling of supercapacitors and batteries will be achieved from first principles to fully understand the dynamic responses of the energy storage components. Then, an adaptive real-time fuzzy neural controller, which is dynamically tuned according to measured and forecast load conditions, will be designed and implemented to optimally manage the power split between the supercapacitor and battery. The research results will not only be validated and evaluated in a lab environment but also tested in real-world conditions with the support of industry partners Averge Technologies (Pty) Ltd and Green Scooter (Pty) Ltd.

  • image

    AI-based geotechnical data analysis

    Funded by Innovate UK and Cathie

    Development of a new service to automate data analysis for seabed engineering developments, driven by AI-enabled predictions from geotechnical data.

  • image

    AI-based professional health and wellbeing support

    Funded by Innovate UK and ART

    Development of system to measure and support employee health and wellbeing, incorporating automated data analysis and full remote functionality, providing validated insights for employer decisions in new and traditional work environments, incluidng post-Covid-19.

  • image

    A Decentralised Secure and Privacy-Preserving E-government System

    Funded by Commonwealth Scholarships Commission

    Electronic Government (e-Government) digitises and innovates public services to businesses, citizens, agencies, employees and other shareholders by utilising Information and Communication Technologies. E-government systems inevitably involves finance, personal, security and other sensitive information, and therefore become the target of cyber attacks through various means, such as malware, spyware, virus, denial of service attacks (DoS), and distributed DoS (DDoS). Despite the protection measures, such as authentication, authorisation, encryption, and firewalls, existing e-Government systems such as websites and electronic identity management systems (eIDs) often face potential privacy issues, security vulnerabilities and suffer from single point of failure due to centralised services. This is getting more challenging along with the dramatically increasing users and usage of e-Government systems due to the proliferation of technologies such as smart cities, internet of things (IoTs), cloud computing and interconnected networks. Thus, there is a need of developing a decentralised secure e-Government system equipped with anomaly detection to enforce system reliability, security and privacy. This PhD work develops a decentralised secure and privacy-preserving e-Government system by innovatively using blockchain technology. Blockchain technology enables the implementation of highly secure and privacy preserving decentralised applications where information is not under the control of any centralised third party. The developed secure and decentralised e-Government system is based on the consortium type of blockchain technology, which is a semi-public and decentralised blockchain system consisting of a group of pre-selected entities or organisations in charge of consensus and decisions making for the benefit of the whole network of peers. Ethereum blockchain solution was used in this project to simulate and validate the proposed system since it is open source and supports off-chain data storage such as images, PDFs, DOCs, contracts, and other files that are too large to be stored in the blockchain or that are required to be deleted or changed in the future, which are essential part of e-Government systems. This PhD work also develops an intrusion detection system (IDS) based on the Dendritic cell algorithm (DCA) for detecting unwanted internal and external traffics to support the proposed blockchain-based e-Government system, because the blockchain database is append-only and immutable. The IDS effectively prevent unwanted transactions such as virus, malware or spyware from being added to the blockchain-based e-Government network. Briefly, the DCA is a class of artificial immune systems (AIS) which was introduce for anomaly detection in computer networks and has beneficial properties such as self-organisation, scalability, decentralised control and adaptability. Three significant improvements have been implemented for DCA-based IDS. Firstly, a new parameters optimisation approach for the DCA is implemented by using the Genetic algorithm (GA). Secondly, fuzzy inference systems approach is developed to solve nonlinear relationship that exist between features during the pre processing stage of the DCA so as to further enhance its anomaly detection performance in e-Government systems. In addition, a multiclass DCA capable of detection multiple attacks is developed in this project, given that the original DCA is a binary classifier and many practical classification problems including computer network intrusion detection datasets are often associated with multiple classes. The effectiveness of the proposed approaches in enforcing security and privacy in e- Government systems are demonstrated through three real-world applications: privacy and integrity protection of information in e Government systems, internal threats detection, and external threats detection. Privacy and integrity protection of information in the proposed e- Government systems is provided by using encryption and validation mechanism offered by the blockchain technology. Experiments demonstrated the performance of the proposed system, and thus its suitability in enhancing security and privacy of information in e-Government systems. The applicability and performance of the DCA-based IDS in e Government systems were examined by using publicly accessible insider and external threat datasets with real world attacks. The results show that, the proposed system can mitigate insider and external threats in e-Government systems whilst simultaneously preserving information security and privacy. The proposed system also could potentially increase the trust and accountability of public sectors due to the transparency and efficiency which are offered by the blockchain applications.

  • image

    AI and Big Data for Cybersecurity

    RAEng and TRF jointly funded project, collaborated with MFL University (Thailand), MoD (UK and Thailand), Thailand Royal Air Force, T-NAT company (Thailand)

    This project aims to develop a new fuzzy modelling approach using the curvature values. This PhD project is co-funded by Yobo Technology Ltd., China and the university. UK Principle Investigator: Dr. Longzhi Yang, and Thailand Principle Investigator: Dr. Tossapon Boongoen

  • Development of a new capability in digital business to grow and diversify income through design and implementation of the SportsAid Athlete Monitoring System (SAMS) to manage athlete performance, health and well-being to win new business in sport and education organisations and with individuals. Please click here to read more about this proejct.

  • image

    Academic Centre of Excellence in Cyber Security Research

    Funded by the National Cyber Security Centre and the Engineering and Physical Sciences Research Council

    PI: Prof. Lynne Coventry, I'm one of the Co-I
    The Academic Centres of Excellence in Cyber Security Research (ACEs-CSR) scheme is sponsored through the Department for Digital, Culture, Media and Sport. It is one of a number of initiatives outlined in the UK Government's £1.9 billion National Cyber Security Strategy 2016-2021, 'Protecting and Promoting the UK in a Digital World', which outlines how Government is working with academia and industry to make the UK more resilient to cyber attacks.
    After successfully meeting the scheme’s tough requirements, Northumbria has now been recognised as an ACE-CSR. The Northumbria Cyber Security Research Group leads the University’s research across this area. This multi-disciplinary group combines technical research on biometric encryption, wireless sensor networks, web security protocols, and image recognition, with human-centred work on usable security, privacy, trust and behaviour change. The group’s work identifies both the virtual and physical risks associated with connected smart cities and complements other work ongoing at the University relating to the digital living space, which explores the intersection of people, place and technology in the digital and urban environment.

  • image

    AI-based effluent treatment recommendation

    Collaborated with a local company in developing a digital solution for effluent treatment recommendation

    Collaborated with a local company in developing a digital solution for for effluent treatment recommendation

  • image

    AI-based running product recommendation

    Collaborated with a local company in developing a digital solution for running product recommendation

    Collaborated with a local company in developing a digital solution for running product recommendation

  • image

    Curvature-based Rule Base Generation for Fuzzy Modelling

    PhD project, co-funded by Yubo Technology Ltd. China and the University

    This project aims to develop an open cyber security research and practice platform. Prinicple supervisor: Dr. Longzhi Yang

  • image

    Intelligent Home Energy Management

    PhD project, funded by the University.

    Smart meters have been partially deployed to UK homes to support consumers in managing their energy and expenditure intelligently, to help implement the CO2 emission reduction goal. However, the information from smart meters is still not enough for efficient intelligent home appliance control, as the residents’ living behaviour patterns, as well as the surrounding environments also have great influences on the control decisions for home appliances. Despite of the advance in machine learning and expert system technologies, the extraction of decision patterns based on residents’ behaviours and the surrounding environments is still of great challenge as every resident has their own living style and each dwelling has its own unique characteristics.

    Take intelligent home heating systems as an example. In order to accurately and economically control the home heating systems such that a property can always be properly pre-heated when the residents getting home whilst no energy is unnecessarily wasted on heating empty dwellings, the control system should be developed based on a well understating of the information from smart meters (such as electricity supply/demand pattern) and the residents’ living styles, and be timely updated when the supply/demand pattern or the living styles change. However, it is unrealistic for system manufacturer to collect highly personalised data and to consequently produce customised heating controllers for each household, such that the heating controller can make accurate decisions based on the residents’ unique living style in addition to the current electricity supply/demand pattern and surrounding environment.

    This research proposes an algorithm which would enable intelligent home heating system to efficiently learn the residents’ behaviour patterns associated with the information from smart maters in a dynamic and adaptive manner. Consequently, mass-production of intelligent heating controllers is allowed. In particular, these devices are initialised by the most general and common rules which are suitable for the majority of people. Then the intelligent controller will learn more customised details in real time after it is deployed. Also, once the electricity supply/demand pattern and the living style or the users of the controller have changed, the controller will be able to catch the changes up in real time. It has been reported that around 40% residential energy use is consumed to deliver ‘unused’ energy services. The proposed approach will potentially provide significant help in reducing energy waste and CO2 emission but in the same time improving the quality of life.

  • image

    Manufacturing Production/Process Control, Scheduling and Optimisation

    Industrial project funded by Simpson Group Print Ltd.

    PI: Dr. Longzhi Yang

    Simpson Group is one of the UK’s leading manufactures of posters and 3D displays for promotions. The company has a number of customers we are all very familiar with, such as Next, Morrison's, and Danbenhams. A promotion usually involves multiple display objects, and a display object is typically produced in a sequence of operations. In addition, an operation can be alternatively conducted by different capable machines with different costs. This project developed a software prototype which intelligently schedules the manufacturing production line use CI techniques and thus improves production efficiency.

  • image

    Robotic Writing System

    Collaborative project, the collaborator was funded by National Natural Science Foundation of China

    This project aims to teach robots to write in an analogy to how humban being writes. PI: Dr. Fei Chao at Xiamen University, China

  • image

    Interaction-based Human Motion Analysis and Retrieval

    Collaborative research, the collaborator is funded by the Engineering and Physical Sciences Research Council

    Interaction based on motion analysis. PI: Dr. Hubert P.H. Shum at Nothumbria University

  • image

    Multi‐layer Lattice Model for Real‐Time Dynamic Character Deformation

    Collaborative research, the collaborator was funded by Japanese Ministry of Education, Culture, Sports, Science and Technology (MEXT)

    MEXT Top Global University Project Scholarship: £5,500, Contributing Researcher (PI: Prof. Shigeo Morishima). Received from the Japanese Ministry of Education, Culture, Sports, Science and Technology (MEXT)To support 1 Japanese PhD (Waseda University) working in the UK for 6 months .

Filter by type:

Sort by year:

Low-Cost Inertial Measurement Unit Calibration With Nonlinear Scale Factors

X. Zhang, C. Zhou, F. Chao, C.M. Lin, L. Yang, C. Shang and Q. Shen
JournalIEEE Transactions on Industrial Informatics, vol. 18, no. 2, pp. 1028 - 1038, 2022.

Abstract

Inertial measurement units (IMUs) have been widely used to provide accurate location and movement measurement solutions, along with the advances of modern manufacturing technologies. The scale factors of accelerometers and gyroscopes are linear when the range of the sensors are reasonably small, but the factor becomes nonlinear when the range gets much bigger. Based on this observation, this article presents a calibration method for low-cost IMU by effectively deriving the nonlinear scale factors of the sensors. Two motion patterns of the sensor on a rigid object are moved to collect data for calibration: One motion pattern is to upcast and rotate the rigid object, and another pattern is to place the rigid object on a stable base in different attitudes. The rotation motion produces centripetal and Coriolis force, which increases the measurement range of accelerometers. Four cost functions with different weight factors and two sets of data are utilized to optimize the IMU parameters. The weight factor comes from derived formula with input values which are the variance of the noise of the sampled data. The proposed approach was validated and evaluated on both synthetic and real-world data sets, and the experimental results demonstrated the superiority of the proposed approach in improving the accuracy of IMU for long-range use. In particular, the errors of acceleration and angular velocity led by our algorithm are significantly smaller than those resulted from the existing approaches using the same testing data sets, demonstrating a remarkable improvement of 64.12% and 47.90%, respectively.

Decoder Choice Network for Metalearning

J. Liu, F. Chao, L. Yang, C.M. Lin, C. Shang, Q. Shen
Journal IEEE Transactions on Cybernetics, 2021.

Abstract

Metalearning has been widely applied for implementing few-shot learning and fast model adaptation. Particularly, existing metalearning methods have been exploited to learn the control mechanism for gradient descent processes, in an effort to facilitate gradient-based learning in gaining high speed and generalization ability. This article presents a novel method that controls the gradient descent process of the model parameters in a neural network, by limiting the model parameters within a low-dimensional latent space. The main challenge for implementing this idea is that a decoder with many parameters may be required. To tackle this problem, the article provides an alternative design of the decoder with a structure that shares certain weights, thereby reducing the number of required parameters. In addition, this work combines ensemble learning with the proposed approach to improve the overall learning performance. Systematic experimental studies demonstrate that the proposed approach offers results superior to the state of the art in performing the Omniglot classification and miniImageNet classification tasks.

A Developmental Evolutionary Learning Framework for Robotic Chinese Stroke Writing

R. Wu, F. Chao, C. Zhou, Y. Huang, L. Yang, C.M. Lin, X. Chang, Q. Shen, C. Shang
Journal IEEE Transactions on Cognitive and Developmental Systems, 2021.

Abstract

The ability of robots to write Chinese strokes, which is recognized as a sophisticated task, involves complicated kinematic control algorithms. The conventional approaches for robotic writing of Chinese strokes often suffer from limited font generation methods, which limits the ability of robots to perform high-quality writing. This paper instead proposes a developmental evolutionary learning framework that enables a robot to learn to write fundamental Chinese strokes. The framework first considers the learning process of robotic writing as an evolutionary easy-to-difficult procedure. Then, a developmental learning mechanism called “Lift-constraint, act and saturate” that stems from developmental robotics is used to determine how the robot learns tasks ranging from simple to difficult by building on the learning results from the easy tasks. The developmental constraints, which include altitude adjustments, number of mutation points, and stroke trajectory points, determine the learning complexity of robot writing. The developmental algorithm divides the evolutionary procedure into three developmental learning stages. In each stage, the stroke trajectory points gradually increase, while the number of mutation points and adjustment altitudes gradually decrease, allowing the learning difficulties involved in these three stages to be categorized as easy, medium, and difficult. Our robot starts with an easy learning task and then gradually progresses to the medium and difficult tasks. Under various developmental constraint setups in each stage, the robot applies an evolutionary algorithm to handle the basic shapes of the Chinese strokes and eventually acquires the ability to write with good quality. The experimental results demonstrate that the proposed framework allows a calligraphic robot to gradually learn to write five fundamental Chinese strokes and also reveal a developmental pattern similar to that of humans. Compared to an evolutionary algorithm without the developmental mechanism, the proposed framework achieves good writing quality more rapidly.

Exclusive lasso-based k-nearest-neighbor classification

L. Qiu, Y. Qu, C. Shang, L. Yang, F. Chao, Q. Shen
Journal Neural Computing and Applications, vol. 33, pp. 14247 - 14261, 2021.

Abstract

Conventionally, the k nearest-neighbor (kNN) classification is implemented with the use of the Euclidean distance-based measures, which are mainly the one-to-one similarity relationships such as to lose the connections between different samples. As a strategy to alleviate this issue, the coefficients coded by sparse representation have played a role of similarity gauger for nearest-neighbor classification as well. Although SR coefficients enjoy remarkable discrimination nature as a one-to-many relationship, it carries out variable selection at the individual level so that possible inherent group structure is ignored. In order to make the most of information implied in the group structure, this paper employs the exclusive lasso strategy to perform the similarity evaluation in two novel nearest-neighbor classification methods. Experimental results on both benchmark data sets and the face recognition problem demonstrate that the EL-based kNN method outperforms certain state-of-the-art classification techniques and existing representation-based nearest-neighbor approaches, in terms of both the size of feature reduction and the classification accuracy.

Embedded YARA rules: strengthening YARA rules utilising fuzzy hashing and fuzzy rules for malware analysis

N. Naik, P. Jenkins, N. Savage, L. Yang, T. Boongoen, N. Iam-On, K. Naik, J. Song
Journal Complex & Intelligent Systems, vol. 7, no. 2, pp. 687 - 702, 2021.

Abstract

The YARA rules technique is used in cybersecurity to scan for malware, often in its default form, where rules are created either manually or automatically. Creating YARA rules that enable analysts to label files as suspected malware is a highly technical skill, requiring expertise in cybersecurity. Therefore, in cases where rules are either created manually or automatically, it is desirable to improve both the performance and detection outcomes of the process. In this paper, two methods are proposed utilising the techniques of fuzzy hashing and fuzzy rules, to increase the effectiveness of YARA rules without escalating the complexity and overheads associated with YARA rules. The first proposed method utilises fuzzy hashing referred to as enhanced YARA rules in this paper, where if existing YARA rules fails to detect the inspected file as malware, then it is subjected to fuzzy hashing to assess whether this technique would identify it as malware. The second proposed technique called embedded YARA rules utilises fuzzy hashing and fuzzy rules to improve the outcomes further. Fuzzy rules countenance circumstances where data are imprecise or uncertain, generating a probabilistic outcome indicating the likelihood of whether a file is malware or not. The paper discusses the success of the proposed enhanced YARA rules and embedded YARA rules through several experiments on the collected malware and goodware corpus and their comparative evaluation against YARA rules.

A computational intelligence enabled honeypot for chasing ghosts in the wires

N. Naik, P. Jenkins, N. Savage, and L. Yang
Journal Complex and Intelligent Systems, vol. 7, no. 1, pp. 477 - 494, 2021.

Abstract

CA honeypot is a concealed security system that functions as a decoy to entice cyberattackers to reveal their information. Therefore, it is essential to disguise its identity to ensure its successful operation. Nonetheless, cyberattackers frequently attempt to uncover these honeypots; one of the most effective techniques for revealing their identity is a fingerprinting attack. Once identified, a honeypot can be exploited as a zombie by an attacker to attack others. Several effective techniques are available to prevent a fingerprinting attack, however, that would be contrary to the purpose of a honeypot, which is designed to interact with attackers to attempt to discover information relating to them. A technique to discover any attempted fingerprinting attack is highly desirable, for honeypots, while interacting with cyberattackers. Unfortunately, no specific method is available to detect and predict an attempted fingerprinting attack in real-time due to the difficulty of isolating it from other attacks. This paper presents a computational intelligence enabled honeypot that is capable of discovering and predicting an attempted fingerprinting attack by using a Principal components analysis and Fuzzy inference system. This proposed system is successfully tested against the five popular fingerprinting tools Nmap, Xprobe2, NetScanTools Pro, SinFP3 and Nessus.

Modality independent adversarial network for generalized zero shot image classification

H. Zhang, Y. Wang, Y. Long, L. Yang, and L. Shao
Journal Neural Networks, vol. 134, pp. 11-22, 2021.

Abstract

Zero Shot Learning (ZSL) aims to classify images of unseen target classes by transferring knowledge from source classes through semantic embeddings. The core of ZSL research is to embed both visual representation of object instance and semantic description of object class into a joint latent space and learn cross-modal (visual and semantic) latent representations. However, the learned representations by existing efforts often fail to fully capture the underlying cross-modal semantic consistency, and some of the representations are very similar and less discriminative. To circumvent these issues, in this paper, we propose a novel deep framework, called Modality Independent Adversarial Network (MIANet) for Generalized Zero Shot Learning (GZSL), which is an end-to-end deep architecture with three submodules. First, both visual feature and semantic description are embedded into a latent hyper-spherical space, where two orthogonal constraints are employed to ensure the learned latent representations discriminative. Second, a modality adversarial submodule is employed to make the latent representations independent of modalities to make the shared representations grab more cross-modal high-level semantic information during training. Third, a cross reconstruction submodule is proposed to reconstruct latent representations into the counterparts instead of themselves to make them capture more modality irrelevant information. Comprehensive experiments on five widely used benchmark datasets are conducted on both GZSL and standard ZSL settings, and the results show the effectiveness of our proposed method.

Automatic stroke generation for style-oriented robotic Chinese calligraphy

G. Lin, Z. Guo, F. Chao, L. Yang, X. Chang, C.-M. Lin, C. Zhou, V. Vijayakumar, and C. Shang
Journal Future Generation Computer Systems, vol. 119, pp. 20-30, 2021.

Abstract

Intelligent robots, as an important type of Cyber-Physical systems, have promising potential to take the central stage in the development of the next-generation of efficient smart systems. Robotic calligraphy is such an attempt, and the current research focuses on the control algorithms of the robotic arms, which usually suffers from significant human inputs and limited writing styles. This paper presents an autonomous robotic writing system for Chinese calligraphy empowered by the proposed automatic stroke matching and generation mechanisms. Thanks to these mechanisms, the robot is able to effectively learn to write any Chinese characters in a style that is sampled by a small amount of handwritten Chinese characters with a certain target writing style. This is achieved by firstly disassembling each given Chinese character into individual strokes using the proposed character disassemble method; then, the writing style of the dissembled strokes is learned by a stroke generation module, which is built upon a generative adversarial learning model. From this, the robot can apply the learned writing style to any Chinese character from a given database, by dissembling the character and then generating the stroke trajectories based on the learned writing style. The experiments confirm the effectiveness of the proposed system in learning writing a certain style of characters based on a small style dataset, as evidenced by the high similarity between the robotic writing results and the handwritten ones according to the Fréchet Inception Distance.

Visual-guided robotic object grasping using dual neural network controllers

W. Fang, F. Chao, C.-M. Lin, D. Zhou, L. Yang, X. Chang, Q. Shen, and C. Shang,
Journal IEEE Transactions on Industrial Informatics, vol. 17, pp. 2282-2291, 2021.

Abstract

It has been a challenging task for a robotic arm to accurately reach and grasp objects, which has drawn much research attention. This article proposes a robotic hand-eye coordination system by simulating the human behavior pattern to achieve a fast and robust reaching ability. This is achieved by two neural-network-based controllers, including a rough reaching movement controller implemented by a pretrained radial basis function for rough reaching movements, and a correction movement controller built from a specifically designed brain emotional nesting network (BENN) for smooth correction movements. In particular, the proposed BENN is designed with high nonlinear mapping ability, with its adaptive laws derived from the Lyapunov stability theorem; from this, the robust tracking performance and accordingly the stability of the proposed control system are guaranteed by the utilization of the H ∞ control approach. The proposed BENN is validated and evaluated by a chaos synchronization simulation, and the overall control system by object grasping tasks through a physical robotic arm in a real-world environment. The experimental results demonstrate the superiority of the proposed control system in reference to those with single neural networks.

Dilated causal convolution with multi-head self attention for sensor human activity recognition

R. Hamad, M. Kimura, L. Yang, W. Woo, B. Wei
Journal Neural Computing and Applications, vol. 33, pp. 13705 - 13722, 2021.

Abstract

ISystems of sensor human activity recognition are becoming increasingly popular in diverse fields such as healthcare and security. Yet, developing such systems poses inherent challenges due to the variations and complexity of human behaviors during the performance of physical activities. Recurrent neural networks, particularly long short-term memory have achieved promising results on numerous sequential learning problems, including sensor human activity recognition. However, parallelization is inhibited in recurrent networks due to sequential operation and computation that lead to slow training, occupying more memory and hard convergence. One-dimensional convolutional neural network processes input temporal sequential batches independently that lead to effectively executed operations in parallel. Despite that, a one-dimensional Convolutional Neural Network is not sensitive to the order of the time steps which is crucial for accurate and robust systems of sensor human activity recognition. To address this problem, we propose a network architecture based on dilated causal convolution and multi-head self-attention mechanisms that entirely dispense recurrent architectures to make efficient computation and maintain the ordering of the time steps. The proposed method is evaluated for human activities using smart home binary sensors data and wearable sensor data. Results of conducted extensive experiments on eight public and benchmark HAR data sets show that the proposed network outperforms the state-of-the-art models based on recurrent settings and temporal models.

A recurrent wavelet-based brain emotional learning network controller for nonlinear systems

J Zhang, F Chao, H Zeng, CM Lin, L Yang
Journal Soft Computing, vol. 26, no. 6, pp. 3013-3028, 2021.

Abstract

Conventional control systems often suffer from the coexistence of nonlinearity and uncertainty. This paper proposes a novel brain emotional neural network to support addressing such challenges. The proposed network integrates a wavelet neural network into a conventional brain emotional learning network. This is further enhanced by the introduction of a recurrent structure to employ the two networks as the two channels of the brain emotional learning network. The proposed network therefore combines the advantages of the wavelet function, the recurrent mechanism, and the brain emotional learning system, for optimal performance on nonlinear problems under uncertain environments. The proposed network works with a bounding compensator to mimic an ideal controller, and the parameters are updated based on the laws derived from the Lyapunov stability analysis theory. The proposed system was applied to two uncertain nonlinear systems, including a Duffing-Homes chaotic system and a simulated 3-DOF spherical joint robot. The experiments demonstrated that the proposed system outperformed other popular neural-network-based control systems, indicating the superiority of the proposed system.

Job shop planning and scheduling for manufacturers with manual operations

L. Yang, J. Li, F. Chao, P. Hackney, M. Flanagan
Journal Expert Systems, vol. 38, no. 7, 2021.

Abstract

Job shop scheduling systems are widely employed to optimise the efficiency of machine utilisation in the manufacturing industry, by searching the most cost-effective permutation of job operations based on the cost of each operation on each compatible machine and the relations between job operations. Such systems are paralysed when the cost of operations are not predictable led by the involvement of complex manual operations. This paper proposes a new genetic algorithm-based job shop scheduling system by integrating a fuzzy learning and inference subsystem in an effort to address this limitation. In particular, the fuzzy subsystem adaptively estimates the completion time and thus cost of each manual task under different conditions based on a knowledge base that is initialised by domain experts and then constantly updated based on its built-in learning ability and adaptability. The manufacturer of Point of Sale and Point of Purchase products has been utilised in this paper as an example case for both theoretical discussion and experimental study. The experimental results demonstrate the promising of the proposed system in improving the efficiency of manual manufacturing operations.

Fuzzy-import hashing: A static analysis technique for malware detection

N Naik, P Jenkins, S Nick, L Yang, B Tossapon, Natthakan, Iam-On
Journal Forensic Science International: Digital Investigation, vol. 37, pp. 301139, 2021.

Abstract

The advent of new malware types and their attack vectors poses serious challenges for security experts in discovering effective malware detection and analysis techniques. The preliminary step in malware analysis is filtering out samples of counterfeit malware from the suspicious samples by classifying them into most likely and unlikely malware categories. This will enable effective utilisation of resources and expertise for the most likely category of samples in subsequent stages and avoid nugatory effort. This process requires a very fast and resource-optimised method as it is applied on a large sample size. Fuzzy hashing and import hashing methods satisfy these requirements of malware analysis, though, with some limitations. Therefore, the proper integration of these methods, may overcome some of the limitations and improve the detection accuracy without affecting the overall performance of analysis. Hence, this paper proposes a fuzzy-import hashing technique, which is the integration of two methods, namely, fuzzy hashing and import hashing. This integration can offer several benefits such as an improved detection rate by complementing each other when one method cannot detect malware, then the other method can; and the generation of fuzzfied results for subsequent clustering or classification, as the import hashing result can be easily merged with the fuzzy hashing result. The success of this proposed fuzzy-import hashing method is demonstrated through several experiments namely: on the collected malware and goodware corpus; a comparative evaluation against the established YARA rules and application in fuzzy c-means clustering.

Adopting Online Teaching and Learning Utilizing AI Technology Enhancements Throughout the COVID-19 Pandemic and Beyond

P. Jenkins, N. Naik, L. Yang
Conference 29th International Conference on Computers in Education, pp. 102-108, 2021

Abstract

Since 2019 the Covid-19 Pandemic has had a significant impact on the daily lives of the populations of the world. Education was no exception, this was a sharp change to online teaching, which was the only available option to keep the education system working. At that time, the education sector was rapidly adapting and attempting to use a variety of online tools, such as Microsoft Teams, Blackboard Collaborate Ultra and Zoom to name a few. Therefore, using these tools for online teaching and learning required significant operational change, that had to happen at pace, as current teaching and learning was scheduled for Face-to-Face (F2F). In addition, greater use was made of Learning Content Management Systems (LCMSs) such as Moodle and Blackboard, which was an existing set of tools. Contemporaneously, these tools were adapting to the teaching and learning by adding additional features, based upon feedback for users. This paper examines some experiences of delivering higher education courses during the COVID-19 pandemic, where it examines the online teaching and learning tools Microsoft Teams and Blackboard Collaborate Ultra, and explores how AI can be used to enhance the process of t

Ankle Variable Impedance Control for Humanoid Robot Upright Balance Control

K. Yin, Y. Xue, Y. Wang, L. Yang
Conference 20th UK Workshop on Computational Intelligence, pp. 203-214, 2021

Abstract

Upright balance control is the most fundamental, yet essential, function of a humanoid robot to enable the performance of various tasks that are traditionally performed by human being in various unstructured environments. Such control schemes were conventionally implemented by developing accurate physical and kinematic models based on fixed torque-ankle states, which often lack robustness to external disturbing forces. This paper presents a variable impedance control method that generates the desired torques for stable humanoid robot upright balance control, to address this limitation. The robustness of the proposed method was brought by a variable parameter approach with the support of the impedance model. The variable parameter of the ankle angle is able to describe the balance state of a humanoid robot, and the proper adjustment of such parameter ensures the effectiveness of the control model. The proposed approach was applied to a humanoid robot on a moving vehicle, and the experimental results demonstrated its efficacy and robustness.

Special issue on emerging trends, challenges and applications in cloud computing

Longzhi Yang, Vijayakumar Varadarajan, Tossapon Boongoen, Nitin Naik
Editorial Wireless Networks, 2021

Abstract

Cloud computing enables ubiquitous and efficient on-demand access to information, data, and computational resources with the support of modern wired and wireless communication technologies. Cloud computing has been very widely used in education, autonomous vehicles, smart cities/homes, renewable energy, healthcare, engineering, business, and telecommunications, amongst others, with the support of the advances in Artificial Intelligence, Internet of Things (IoT), and Data Science. Such technologies and their applications have made significant impact to the way people live and do business by offering online services and instant communications.

Special issue on recent advances in data science and systems

L Yang, J Hu, CL Hung
Editorial Expert Systems, vol. 38, no. 3, e12735, 2021

Abstract

As an interdisciplinary area, Data Science draws scientific inquiry from a broad range of subject areas such as statistics, mathematics, computer science, machine learning, optimisation, signal processing, information retrieval, databases, cloud computing, computer vision, natural language processing, and so forth. Data Science aims to deliver valuable insights from data, and to meet the challenges of processing very large datasets, that is, Big Data, with new data continuously generated from various channels, such as smart devices, web, mobile and social media.

Error Controlled Actor-Critic Method to Reinforcement Learning

X Gao, F Chao, C Zhou, Z Ge, CM Lin, L Yang, X Chang, C Shang
Preprint arXiv preprint arXiv:2109.02517, 2021

Abstract

On error of value function inevitably causes an overestimation phenomenon and has a negative impact on the convergence of the algorithms. To mitigate the negative effects of the approximation error, we propose Error Controlled Actor-critic which ensures confining the approximation error in value function. We present an analysis of how the approximation error can hinder the optimization process of actor-critic methods.Then, we derive an upper boundary of the approximation error of Q function approximator and find that the error can be lowered by restricting on the KL-divergence between every two consecutive policies when training the policy. The results of experiments on a range of continuous control tasks demonstrate that the proposed actor-critic algorithm apparently reduces the approximation error and significantly outperforms other model-free RL algorithms.

GANCCRobot: Generative adversarial nets based Chinese calligraphy robot

R. Wu, C. Zhou, F. Chao, L. Yang, C.-M. Lin, and C. Shang
Journal Information Sciences, vol. 516, pp. 474-490, 2020.

Abstract

Robotic calligraphy, as a typical application of robot movement planning, is of great significance for the inheritance and education of calligraphy culture. The existing implementations of such robots often suffer from its limited ability for font generation and evaluation, leading to poor writing style diversity and writing quality. This paper proposes a calligraphic robotic framework based on the generative adversarial nets (GAN) to address such limitation. The robot implemented using such framework is able to learn to write fundamental Chinese character strokes with rich diversities and good quality that is close to the human level, without the requirement of specifically designed evaluation functions thanks to the employment of the revised GAN. In particular, the type information of the stroke is introduced as condition information, and the latent codes are applied to maximize the style quality of the generated strokes. Experimental results demonstrate that the proposed model enables a calligraphic robot to successfully write fundamental Chinese strokes based on a given type and style, with overall good quality. Although the proposed model was evaluated in this report using calligraphy writing, the underpinning research is readily applicable to many other applications, such as robotic graffiti and character style conversion.

Type-2 fuzzy hybrid controller network for robotic systems

F. Chao, D. Zhou, C. Lin, L. Yang, C. Zhou, and C. Shang
Journal IEEE Transactions on Cybernetics, vol. 50, pp. 3778-3792, 2020.

Abstract

Dynamic control, including robotic control, faces both the theoretical challenge of obtaining accurate system models and the practical difficulty of defining uncertain system bounds. To facilitate such challenges, this paper proposes a control system consisting of a novel type of fuzzy neural network and a robust compensator controller. The new fuzzy neural network is implemented by integrating a number of key components embedded in a Type-2 fuzzy cerebellar model articulation controller (CMAC) and a brain emotional learning controller (BELC) network, thereby mimicking an ideal sliding mode controller. The system inputs are fed into the neural network through a Type-2 fuzzy inference system (T2FIS), with the results subsequently piped into sensory and emotional channels which jointly produce the final outputs of the network. That is, the proposed network estimates the nonlinear equations representing the ideal sliding mode controllers using a powerful compensator controller with the support of T2FIS and BELC, guaranteeing robust tracking of the dynamics of the controlled systems. The adaptive dynamic tuning laws of the network are developed by exploiting the popular brain emotional learning rule and the Lyapunov function. The proposed system was applied to a robot manipulator and a mobile robot, demonstrating its efficacy and potential; and a comparative study with alternatives indicates a significant improvement by the proposed system in performing the intelligent dynamic control.

Artificial human balance control by calf muscle activation modelling

K. Yin, J. Chen, K. Xiang, M. Pang, B. Tang, J. Li and L. Yang
Journal IEEE Access, vol. 8, pp. 86732-86744, 2020.

Abstract

The natural neuromuscular model has greatly inspired the development of control mechanisms in addressing the uncertainty challenges in robotic systems. Although the underpinning neural reaction of posture control remains unknown, recent studies suggest that muscle activation driven by the nervous system plays a key role in human postural responses to environmental disturbance. Given that the human calf is mainly formed by two muscles, this paper presents an integrated calf control model with the two comprising components representing the activations of the two calf muscles. The contributions of each component towards the artificial control of the calf are determined by their weights, which are carefully designed to simulate the natural biological calf. The proposed calf modelling has also been applied to robotic ankle exoskeleton control. The proposed work was validated and evaluated by both biological and engineering simulation approaches, and the experimental results revealed that the proposed model successfully performed over 92% of the muscle activation naturally made by human participants, and the actions led by the simulated ankle exoskeleton wearers were overall consistent with that by the natural biological response.

Integration of an actor-critic model and generative adversarial networks for a Chinese calligraphy robot

R. Wu, C. Zhou, F. Chao, L. Yang, C.-M. Lin, and C. Shang
Journal Neurocomputing, vol. 388, 2020.

Abstract

As a combination of robotic motion planning and Chinese calligraphy culture, robotic calligraphy plays a significant role in the inheritance and education of Chinese calligraphy culture. Most existing calligraphy robots focus on enabling the robots to learn writing through human participation, such as human–robot interactions and manually designed evaluation functions. However, because of the subjectivity of art aesthetics, these existing methods require a large amount of implementation work from human engineers. In addition, the written results cannot be accurately evaluated. To overcome these limitations, in this paper, we propose a robotic calligraphy model that combines a generative adversarial network (GAN) and deep reinforcement learning to enable a calligraphy robot to learn to write Chinese character strokes directly from images captured from Chinese calligraphic textbooks. In our proposed model, to automatically establish an aesthetic evaluation system for Chinese calligraphy, a GAN is first trained to understand and reconstruct stroke images. Then, the discriminator network is independently extracted from the trained GAN and embedded into a variant of the reinforcement learning method, the “actor-critic model”, as a reward function. Thus, a calligraphy robot adopts the improved actor-critic model to learn to write multiple character strokes. The experimental results demonstrate that the proposed model successfully allows a calligraphy robot to write Chinese character strokes based on input stroke images. The performance of our model, compared with the state-of-the-art deep reinforcement learning method, shows the efficacy of the combination approach. In addition, the key technology in this work shows promise as a solution for robotic autonomous assembly.

Interaction-based human activity comparison

Y. Shen, L. Yang, E. S. L. Ho, and H. P. H. Shum
Journal IEEE Transactions on Visualization and Computer Graphics, vol. 26, pp. 2620-2633, 2020.

Abstract

Traditional methods for motion comparison consider features from individual characters. However, the semantic meaning of many human activities is usually defined by the interaction between them, such as a high-five interaction of two characters. There is little success in adapting interaction-based features in activity comparison, as they either do not have a fixed topology or are in high dimensional. In this paper, we propose a unified framework for activity comparison from the interaction point of view. Our new metric evaluates the similarity of interaction by adapting the Earth Mover's Distance onto a customized geometric mesh structure that represents spatial-temporal interactions. This allows us to compare different classes of interactions and discover their intrinsic semantic similarity. We created five interaction databases of different natures, covering both two-characters (synthetic and real-people) and character-object interactions, which are open for public uses. We demonstrate how the proposed metric aligns well with the semantic meaning of the interaction. We also apply the metric in interaction retrieval and show how it outperforms existing ones. The proposed method can be used for unsupervised activity detection in monitoring systems and activity retrieval in smart animation systems.

Semantic combined network for zero‐shot scene parsing

Y Wang, H Zhang, S Wang, Y Long, L Yang
Journal IET Image Processing, vol. 14, no. 4, pp. 757-765, 2020.

Abstract

Recently, image-based scene parsing has attracted increasing attention due to its wide application. However, conventional models can only be valid on images with the same domain of the training set and are typically trained using discrete and meaningless labels. Inspired by the traditional zero-shot learning methods which employ auxiliary side information to bridge the source and target domains, the authors propose a novel framework called semantic combined network (SCN), which aims at learning a scene parsing model only from the images of the seen classes while targeting on the unseen ones. In addition, with the assistance of semantic embeddings of classes, the proposed SCN can further improve the performances of traditional fully supervised scene parsing methods. Extensive experiments are conducted on the data set Cityscapes, and the results show that the proposed SCN can perform well on both zero-shot scene parsing (ZSSP) and generalised ZSSP settings based on several state-of-the-art scenes parsing architectures. Furthermore, the authors test the proposed model under the traditional fully supervised setting and the results show that the proposed SCN can also significantly improve the performances of the original network models.

An LSTM Based Generative Adversarial Architecture for Robotic Calligraphy Learning System

F Chao, G Lin, L Zheng, X Chang, CM Lin, L Yang, C Shang
Journal Sustainability, vo. 12, no. 21, pp. 9092, 2020.

Abstract

Robotic calligraphy is a very challenging task for the robotic manipulators, which can sustain industrial manufacturing. The active mechanism of writing robots require a large sized training set including sequence information of the writing trajectory. However, manual labelling work on those training data may cause the time wasting for researchers. This paper proposes a machine calligraphy learning system using a Long Short-Term Memory (LSTM) network and a generative adversarial network (GAN), which enables the robots to learn and generate the sequences of Chinese character stroke (i.e., writing trajectory). In order to reduce the size of the training set, a generative adversarial architecture combining an LSTM network and a discrimination network is established for a robotic manipulator to learn the Chinese calligraphy regarding its strokes. In particular, this learning system converts Chinese character stroke image into the trajectory sequences in the absence of the stroke trajectory writing sequence information. Due to its powerful learning ability in handling motion sequences, the LSTM network is used to explore the trajectory point writing sequences. Each generation process of the generative adversarial architecture contains a number of loops of LSTM. In each loop, the robot continues to write by following a new trajectory point, which is generated by LSTM according to the previously written strokes. The written stroke in an image format is taken as input to the next loop of the LSTM network until the complete stroke is finally written. Then, the final output of the LSTM network is evaluated by the discriminative network. In addition, a policy gradient algorithm based on reinforcement learning is employed to aid the robot to find the best policy. The experimental results show that the proposed learning system can effectively produce a variety of high-quality Chinese stroke writing.

Joint learning of temporal models to handle imbalanced data for human activity recognition

RA Hamad, L Yang, WL Woo, B Wei
Journal Applied Sciences, vol. 10, no. 15, pp. 5293, 2020.

Abstract

Human activity recognition has become essential to a wide range of applications, such as smart home monitoring, health-care, surveillance. However, it is challenging to deliver a sufficiently robust human activity recognition system from raw sensor data with noise in a smart environment setting. Moreover, imbalanced human activity datasets with less frequent activities create extra challenges for accurate activity recognition. Deep learning algorithms have achieved promising results on balanced datasets, but their performance on imbalanced datasets without explicit algorithm design cannot be promised. Therefore, we aim to realise an activity recognition system using multi-modal sensors to address the issue of class imbalance in deep learning and improve recognition accuracy. This paper proposes a joint diverse temporal learning framework using Long Short Term Memory and one-dimensional Convolutional Neural Network models to improve human activity recognition, especially for less represented activities. We extensively evaluate the proposed method for Activities of Daily Living recognition using binary sensors dataset. A comparative study on five smart home datasets demonstrate that our proposed approach outperforms the existing individual temporal models and their hybridization. Furthermore, this is particularly the case for minority classes in addition to reasonable improvement on the majority classes of human activities.

Anomaly Detection for Internet of Things (IoT) Using an Artificial Immune System

N Elisa, L Yang, F Chao, N Naik
Conference Proceedings of the 12th International Conference on Soft Computing and Pattern Recognition (SoCPaR 2020), pp. 858-867, 2020

Abstract

Internet of Things (IoT) have demonstrated significant impact on all aspects of human daily lives due to their pervasive applications in areas such as telehealth, home appliances, surveillance, and wearable devices. The number of IoT devices and sensors connected to the Internet across the world is expected to reach over 50 billion by the end of 2020. The connection of such rapidly increasing number of IoT devices to the Internet leads to concerns in cyber-attacks such as malware, worms, denial of service attack (DoS) and distributed DoS attack (DDoS). To prevent these attacks from compromising the performance of IoT devices, various approaches for detecting and mitigating cyber security threats have been developed. This paper reports an IoT attack and anomaly detection approach by using the dendritic cell algorithm (DCA). In particular, DCA is an artificial immune system (AIS), which is developed from the inspiration of the working principles and characteristic behaviours of the human immune system (HIS), specifically for the purpose of detecting anomalies in networked systems. The performance of the DCA on detecting IoT attacks is evaluated using publicly available IoT datasets, including DoS, DDoS, Reconnaissance, Keylogging, and Data exfiltration. The experimental results show that, the DCA achieved a comparable detection performance to some of the commonly used classifiers, such as decision trees, random forests, support vector machines, artificial neural network and naïve Bayes, but with reasonably high computational efficiency.

Fuzzy hashing aided enhanced YARA rules for malware triaging

N. Naik, P. Jenkins, N. Savage, L. Yang, K. Naik, J. Song, T. Boongoen, N. Iam-On
Conference 2020 IEEE Symposium Series on Computational Intelligence (SSCI), 2020

Abstract

Cybercriminals are becoming more sophisticated wearing a mask of anonymity and unleashing more destructive malware on a daily basis. The biggest challenge is coping with the abundance of malware created and filtering targeted samples of destructive malware for further investigation and analysis whilst discarding any inert samples, thus optimising the analysis by saving time, effort and resources. The most common technique is malware triaging to separate likely malware and unlikely malware samples. One such triaging technique is YARA rules, commonly used to detect and classify malware based on string and pattern matching, rules are triggered and alerted when their condition is satisfied. This pattern matching technique used by YARA rules and its detection rate can be improved in several ways, however, it can lead to bulky and complex rules that affect the performance of YARA rules. This paper proposes a fuzzy hashing aided enhanced YARA rules to improve the detection rate of YARA rules without significantly increasing the complexity and overheads inherent in the process. This proposed approach only uses an additional fuzzy hashing alongside basic YARA rules to complement each other, so that when one method cannot detect a match, then the other technique can. This work employs three triaging methods fuzzy hashing, import hashing and YARA rules to perform extensive experiments on the collected malware samples. The detection rate of enhanced YARA rules is compared against the detection rate of the employed triaging methods to demonstrate the improvement in the overall triaging results.

A Novel Self-Organizing Emotional CMAC Network for Robotic Control

J Zhang, Q Li, X Chang, F Chao, CM Lin, L Yang, TT Huynh, L Zheng, C Zhou, C Shang
Conference 2020 International Joint Conference on Neural Networks (IJCNN), 2020

Abstract

This paper proposes a self-organizing control system for uncertain nonlinear systems. The proposed neural network is composed of a conventional brain emotional learning network (BEL) and a cerebellar model articulation controller network (CMAC). The input value of the network is feed to a BEL channel and a CMAC channel. The output of the network is generated by the comprehensive action of the two channels. The structure of the network is dynamic, using a self-organizing algorithm allows increasing or decreasing weight layers. The parameters of the proposed network are on-line tuned by the brain emotional learning rules; the updating rules of CMAC and the robust controller are derived from the Lyapunov function; in addition, stability analysis theory is used to guaranty the proposed controller's convergence. A simulated mobile robot is applied to prove the effectiveness of the proposed control system. By comparing with the performance of other neural-network-based control systems, the proposed network produces better performance.

Embedding fuzzy rules with YARA rules for performance optimisation of malware analysis

N Naik, P Jenkins, N Savage, L Yang, K Naik, J Song
Conference 2020 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), 2020

Abstract

YARA rules utilises string or pattern matching to perform malware analysis and is one of the most effective methods in use today. However, its effectiveness is dependent on the quality and quantity of YARA rules employed in the analysis. This can be managed through the rule optimisation process, although, this may not necessarily guarantee effective utilisation of YARA rules and its generated findings during its execution phase, as the main focus of YARA rules is in determining whether to trigger a rule or not, for a suspect sample after examining its rule condition. YARA rule conditions are Boolean expressions, mostly focused on the binary outcome of the malware analysis, which may limit the optimised use of YARA rules and its findings despite generating significant information during the execution phase. Therefore, this paper proposes embedding fuzzy rules with YARA rules to optimise its performance during the execution phase. Fuzzy rules can manage imprecise and incomplete data and encompass a broad range of conditions, which may not be possible in Boolean logic. This embedding may be more advantageous when the YARA rules become more complex, resulting in multiple complex conditions, which may not be processed efficiently utilising Boolean expressions alone, thus compromising effective decision-making. This proposed embedded approach is applied on a collected malware corpus and is tested against the standard and enhanced YARA rules to demonstrate its success.

Fuzzy-Import Hashing: A malware analysis approach

N Naik, P Jenkins, N Savage, L Yang, T Boongoen, N Iam-On
Conference 2020 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), 2020

Abstract

Malware has remained a consistent threat since its emergence, growing into a plethora of types and in large numbers. In recent years, numerous new malware variants have enabled the identification of new attack surfaces and vectors, and have become a major challenge to security experts, driving the enhancement and development of new malware analysis techniques to contain the contagion. One of the preliminary steps of malware analysis is to remove the abundance of counterfeit malware samples from the large collection of suspicious samples. This process assists in the management of man and machine resources effectively in the analysis of both unknown and likely malware samples. Hashing techniques are one of the fastest and efficient techniques for performing this preliminary analysis such as fuzzy hashing and import hashing. However, both hashing methods have their limitations and they may not be effective on their own, instead the combination of two distinctive methods may assist in improving the detection accuracy and overall performance of the analysis. This paper proposes a Fuzzy-Import hashing technique which is the combination of fuzzy hashing and import hashing to improve the detection accuracy and overall performance of malware analysis. This proposed Fuzzy-Import hashing offers several benefits which are demonstrated through the experimentation performed on the collected malware samples and compared against stand-alone techniques of fuzzy hashing and import hashing.

A comparative study of genetic algorithm and particle swarm optimisation for dendritic cell algorithm

N Elisa, L Yang, F Chao, N Naik
Conference 2020 IEEE Congress on Evolutionary Computation (CEC), 2020

Abstract

Dendritic cell algorithm (DCA) is a class of artificial immune systems that was originally developed for anomaly detection in networked systems and later as a general binary classifier. Conventionally, in its life cycle, the DCA goes through four phases including feature categorisation into artificial signals, context detection of data items, context assignment, and finally labeling of data items as either abnormal or normal class. During the context detection phase, the DCA requires users to manually pre-define the parameters used by its weighted function to process the signals and data items. Notice that the manual derivation of the parameters of the DCA cannot guarantee the optimal set of weights being used, research attention has thus been attracted to the optimisation of the parameters. This paper reports a systematic comparative study between Genetic algorithm (GA) and Particle Swarm optimisation (PSO) on parameter optimisation for DCA. In order to evaluate the performance of GADCA and PSO-DCA, twelve publicly available datasets from UCI machine learning repository were employed. The performance results based on the computational time, classification accuracy, sensitivity, F-measure, and precision show that, the GA-DCA overall outperforms PSO-DCA for most of the datasets.

Consortium Blockchain for Security and Privacy-Preserving in E-government Systems

N Elisa, L Yang, H. Li, F Chao, N Naik
Conference Proceedings of The 19th International Conference on Electronic Business (pICEB), pp. 99-107, 2020

Abstract

Since its inception as a solution for secure cryptocurrencies sharing in 2008, the blockchain technology has now become one of the core technologies for secure data sharing and storage over trustless and decentralised peer-to-peer systems. E-government is amongst the systems that stores sensitive information about citizens, businesses and other affiliates, and therefore becomes the target of cyber attackers. The existing e-government systems are centralised and thus subject to single point of failure. This paper proposes a secure and decentralised e-government system based on the consortium blockchain technology, which is a semi-public and decentralised blockchain system consisting of a group of pre-selected entities or organisations in charge of consensus and decisions making for the benefit of the whole network of peers. In addition, a number of e-government nodes are pre-selected to perform the tasks of user and transaction validation before being added to the blockchain network. Accordingly, e-government users of the consortium blockchain network are given the rights to create, submit, access, and review transactions. Performance evaluation on single transaction time and transactions processed per second demonstrate the practicability of the proposed consortium blockchain-based e-government system for secure information sharing amongst all stakeholders.

Resilience and effective learning in first-year undergraduate computer science

T Prickett, J Walters, L Yang, M Harvey, T Crick
Conference Proceedings of the 2020 ACM Conference on Innovation and Technology in Computer Science Education (ITiCSE'20), pp. 19–25, 2020

Abstract

Many factors have been shown to be important for supporting effective learning and teaching -- and thus progression and success -- in higher education. While factors such as key introductory-level (CS1) knowledge and skills, as well as pre-university learning and qualifications, have been extensively explored, the impact of measures of positive psychology are less well understood for the discipline of computer science. University study can be a period of significant transition for many students; therefore an individual's positive psychology may have considerable impact upon their response to these challenges. This work investigates the relationships between effective learning and success (first-year performance and attendance) and two measures of positive psychology: Grit and the Nicolson McBride Resilience Quotient (NMRQ). Data was captured by integrating Grit (N=58) and Resilience (N=50) questionnaires and related coaching into the first-year of the undergraduate computer science programme at a single UK university. Analyses demonstrate that NMRQ is significantly linked to attendance and performance for individual subjects and year average marks; however, this was not the case for Grit. This suggests that development of targeted interventions to support students in further developing their resilience could support their learning, as well as progression and retention. Resilience could be used, in concert with other factors such as learning analytics, to augment a range of existing models to predict future student success, allowing targeted academic and pastoral support.

Exploring resilience for effective learning in computer science education

T Prickett, T Crick, M Harvey, J Walters, L Yang
Conference Cambridge Computing Education Research Symposium, 2020

Abstract

Many factors have been shown to be important for supporting effective learning and teaching – and thus progression and success – in formal educational contexts. While factors such as key introductory-level computer science knowledge and skills, as well as pre-university learning and qualifications, have been extensively explored, the impact of measures of positive psychology are less well understood for the discipline of computer science. This preliminary work investigates the relationships between effective learning and success, and two measures of positive psychology, Grit (Duckworth’s 12-item Grit scale) [6] and the Nicolson McBride Resilience Quotient (NMRQ) [3], in success in first-year undergraduate computer science to provide insight into the factors that impact on the transition from secondary education into tertiary education.

Big data in healthcare: Challenges and promise

S Kale, H Tamakuwala, V Vijayakumar, L Yang, BS Rawal Kshatriya
Conference Proceedings of 6th International Conference on Big Data and Cloud Computing Challenges, pp 3–17, 2020

Abstract

Recently, the growth of the clinical sector and the technologies used in combination with the healthcare sector has resulted in the massive growth of the data that is being produced. To handle, store, and analyze such massive amounts of data, big data techniques are being used in the healthcare sector. This article features the gigantic effects of big data on restorative partners, patients, doctors, pharmaceutical and therapeutic administrators, and healthcare backup plans, and furthermore audits the various difficulties that must be considered to get the best benefits from this big data and accessible applications.

Histogram of fuzzy local spatio-temporal descriptors for video action recognition

Z. Zuo, L. Yang, Y. Liu, F. Chao, R. Song, and Y. Qu
Journal IEEE Transactions on Industrial Informatics, vol. 16, pp. 4059-4067, 2019.

Abstract

Feature extraction plays a vital role in visual action recognition. Many existing gradient-based feature extractors, including histogram of oriented gradients, histogram of optical flow, motion boundary histograms, and histogram of motion gradients, build histograms for representing different actions over the spatio-temporal domain in a video. However, these methods require to set the number of bins for information aggregation in advance. Varying numbers of bins usually lead to inherent uncertainty within the process of pixel voting with regard to the bins in the histogram. This article proposes a novel method to handle such uncertainty by fuzzifying these feature extractors. The proposed approach has two advantages: it better represents the ambiguous boundaries between the bins and, thus, the fuzziness of the spatio-temporal visual information entailed in videos; and the contribution of each pixel is flexibly controlled by a fuzziness parameter for various scenarios. The proposed family of fuzzy descriptors and a combination of them are evaluated on two publicly available datasets, demonstrating that the proposed approach outperforms the original counterparts and other state-of-the-art methods.

An early diagnosis of oral cancer based on three-dimensional convolutional neural networks

S. Xu, C. Liu, Y. Zong, S. Chen, Y. Lu, L. Yang, E. Y. K. Ng, Y. Wang, Y. Wang, Y. Liu, W. Hu, and C. Zhang
Journal IEEE Access, vol. 7, pp. 158603-158611, 2019.

Abstract

Three-dimensional convolutional neural networks (3DCNNs), a rapidly evolving modality of deep learning, has gained popularity in many fields. For oral cancers, CT images are traditionally processed using two-dimensional input, without considering information between lesion slices. In this paper, we established a 3DCNNs-based image processing algorithm for the early diagnosis of oral cancers, which was compared with a 2DCNNs-based algorithm. The 3D and 2D CNNs were constructed using the same hierarchical structure to profile oral tumors as benign or malignant. Our results showed that 3DCNNs with dynamic characteristics of the enhancement rate image performed better than 2DCNNS with single enhancement sequence for the discrimination of oral cancer lesions. Our data indicate that spatial features and spatial dynamics extracted from 3DCNNs may inform future design of CT-assisted diagnosis system.

Multi-criterion mammographic risk analysis supported with multi-label fuzzy-rough feature selection

Y. Qu, G. Yue, C. Shang, L. Yang, R. Zwiggelaar, and Q. Shen
Journal Artificial Intelligence in Medicine, vol. 100, p. 101722, 2019.

Abstract

Context and background Breast cancer is one of the most common diseases threatening the human lives globally, requiring effective and early risk analysis for which learning classifiers supported with automated feature selection offer a potential robust solution. Motivation Computer aided risk analysis of breast cancer typically works with a set of extracted mammographic features which may contain significant redundancy and noise, thereby requiring technical developments to improve runtime performance in both computational efficiency and classification accuracy. Hypothesis Use of advanced feature selection methods based on multiple diagnosis criteria may lead to improved results for mammographic risk analysis. Methods An approach for multi-criterion based mammographic risk analysis is proposed, by adapting the recently developed multi-label fuzzy-rough feature selection mechanism. Results A system for multi-criterion mammographic risk analysis is implemented with the aid of multi-label fuzzy-rough feature selection and its performance is positively verified experimentally, in comparison with representative popular mechanisms. Conclusions The novel approach for mammographic risk analysis based on multiple criteria helps improve classification accuracy using selected informative features, without suffering from the redundancy caused by such complex criteria, with the implemented system demonstrating practical efficacy.

A robotic writing framework-learning human aesthetic preferences via human-machine interactions

X. Gao, C. Zhou, F. Chao, L. Yang, C. Lin, and C. Shang
Journal IEEE Access, vol. 7, pp. 144043-144053, 2019.

Abstract

Intelligent robots are required to fully understand human intentions and operations in order to support or collaborate with humans to complete complicated tasks, which is typically implemented by employing human-machine interaction techniques. This paper proposes a new robotic learning framework to perform numeral writing tasks by investigating human-machine interactions with human preferences. In particular, the framework implements a trajectory generative module using a generative adversarial network (GAN)-based method and develops a human preference feedback system to enable the robot to learn human preferences. In addition, a convolutional neural network, acting as a discriminative network, classifies numeral images to support the development of the basic numeral writing ability, and another convolutional neural network, acting as a human preference network, learns a human user's aesthetic preference by taking the feedback on two written numerical images during the training process. The experimental results show that the written numerals based on the preferences of ten users were different from those of the training data set and that the writing models with the preferences from different users generate numerals in different styles, as evidenced by the Fréchet inception distance (FID) scores. The FID scores of the proposed framework with a preference network were noticeably greater than those of the framework without a preference network. This phenomenon indicates that the human-machine interactions effectively guided the robotic system to learn different writing styles. These results prove that the proposed approach is able to enable the calligraphy robot to successfully write numerals in accordance with the preferences of a human user.

A data-driven robotic Chinese calligraphy system using convolutional auto-encoder and differential evolution

X. Gao, C. Zhou, F. Chao, L. Yang, C.-M. Lin, T. Xu, C. Shang, and Q. Shen
Journal Knowledge-Based Systems, vol. 182, pp. 104802, 2019.

Abstract

The Chinese stroke evaluation and generation systems required in an autonomous calligraphy robot play a crucial role in producing high-quality writing results with good diversity. These systems often suffer from inefficiency and non-optima despite of intensive research effort investment by the robotic community. This paper proposes a new learning system to allow a robot to automatically learn to write Chinese calligraphy effectively. In the proposed system, the writing quality evaluation subsystem assesses written strokes using a convolutional auto-encoder network (CAE), which enables the generation of aesthetic strokes with various writing styles. The trained CAE network effectively excludes poorly written strokes through stroke reconstruction, but guarantees the inheritance of information from well-written ones. With the support of the evaluation subsystem, the writing trajectory model generation subsystem is realized by multivariate normal distributions optimized by differential evolution (DE), a type of heuristic optimization search algorithm. The proposed approach was validated and evaluated using a dataset of nine stroke categories; high-quality written strokes have been resulted with good diversity which shows the robustness and efficacy of the proposed approach and its potential in autonomous action-state space exploration for other real-world applications.

A Deep-Learning-Driven Light-Weight Phishing Detection Sensor

B. Wei, R. Hamad, L. Yang, X. He, H. Wang, B. Gao, L. Woo
Journal Sensors, vol. 19, pp. 4258, 2019.

Abstract

This paper designs an accurate and low-cost phishing detection sensor by exploring deep learning techniques. Phishing is a very common social engineering technique. The attackers try to deceive online users by mimicking a uniform resource locator (URL) and a webpage. Traditionally, phishing detection is largely based on manual reports from users. Machine learning techniques have recently been introduced for phishing detection. With the recent rapid development of deep learning techniques, many deep-learning-based recognition methods have also been explored to improve classification performance. This paper proposes a light-weight deep learning algorithm to detect the malicious URLs and enable a real-time and energy-saving phishing detection sensor. Experimental tests and comparisons have been conducted to verify the efficacy of the proposed method. According to the experiments, the true detection rate has been improved. This paper has also verified that the proposed method can run in an energy-saving embedded single board computer in real-time.

Personalised control of robotic ankle exoskeleton through experience-based adaptive fuzzy inference

K. Yin, K. Xiang, M. Pang, J. Chen, P. Anderson, and L. Yang
Journal IEEE Access, vol. 7, 2019.

Abstract

Robotic exoskeletons have emerged as effective rehabilitation and ability-enhancement tools, by mimicking or supporting natural body movements. The control schemes of exoskeletons are conventionally developed based on fixed torque-ankle state relationship or various human models, which are often lack of flexibility and adaptability to accurately address personalized movement assistance needs. This paper presents an adaptive control strategy for personalized robotic ankle exoskeleton in an effort to address this limitation. The adaptation was implemented by applying the experience-based fuzzy rule interpolation approach with the support of a muscle-tendon complex model. In particular, this control system is initialized based on the most common requirements of a “standard human model,” which is then evolved during its performance by effectively using the feedback collected from the wearer to support different body shapes and assistance needs. The experimental results based on different human models with various support demands demonstrate the power of the proposed control system in improving the adaptability, and thus applicability, of robotic ankle exoskeletons.

Use of Automatic Chinese Character Decomposition and Human Gestures for Chinese Calligraphy Robots

F. Chao, Y. Huang, C. Lin, L. Yang, H. Hu and C. Zhou
Journal IEEE Transactions on Human-Machine Systems, vol. 49, no. 1, pp. 47-58, 2019.

Abstract

Conventional Chinese calligraphy robots often suffer from the limited sizes of predefined font databases, which prevent the robots from writing new characters. This paper presents a robotic handwriting system to address such limitations, which extracts Chinese characters from textbooks and uses a robot's manipulator to write the characters in a different style. The key technologies of the proposed approach include the following: 1) automatically decomposing Chinese characters into strokes using Harris corner detection technology and 2) matching the decomposed strokes to robotic writing trajectories learned from human gestures. Briefly, the system first decomposes a given Chinese character into a set of strokes and obtains the stroke trajectory writing ability by following the gestures performed by a human demonstrator. Then, it applies a stroke classification method that recognizes the decomposed strokes as robotic writing trajectories. Finally, the robot arm is driven to follow the trajectories and thus write the Chinese character. Seven common Chinese characters have been used in an experiment for system validation and evaluation. The experimental results demonstrate the power of the proposed system, given that the robot successfully wrote all the testing characters in the given Chinese calligraphic style.

A recurrent emotional CMAC neural network controller for vision-based mobile robots

W. Fang, F. Chao, L. Yang, C.-M. Lin, C. Shang, C. Zhou, and Q. Shen
Journal Neurocomputing, vol. 334, 2019.

Abstract

Vision-based mobile robots often suffer from the difficulties of high nonlinear dynamics and precise positioning requirements, which leads to the development demand of more powerful nonlinear approximation in controlling and monitoring of mobile robots. This paper proposes a recurrent emotional cerebellar model articulation controller (RECMAC) neural network in meeting such demand. In particular, the proposed network integrates a recurrent loop and an emotional learning mechanism into a cerebellar model articulation controller (CMAC), which is implemented as the main component of the controller module of a vision-based mobile robot. Briefly, the controller module consists of a sliding surface, the RECMAC, and a compensator controller. The incorporation of the recurrent structure in a slide model neural network controller ensures the retaining of the previous states of the robot to improve its dynamic mapping ability. The convergence of the proposed system is guaranteed by applying the Lyapunov stability analysis theory. The proposed system was validated and evaluated by both simulation and a practical moving-target tracking task. The experimentation demonstrated that the proposed system outperforms other popular neural network-based control systems, and thus it is superior in approximating highly nonlinear dynamics in controlling vision-based mobile robots.

Towards big data governance in cybersecurity

L Yang, J Li, N Elisa, T Prickett, F Chao
Journal Data-Enabled Discovery and Applications, vol. 3, no. 1, pp. 1-12, 2019.

Abstract

Big data refers to large complex structured or unstructured data sets. Big data technologies enable organisations to generate, collect, manage, analyse, and visualise big data sets, and provide insights to inform diagnosis, prediction, or other decision-making tasks. One of the critical concerns in handling big data is the adoption of appropriate big data governance frameworks to (1) curate big data in a required manner to support quality data access for effective machine learning and (2) ensure the framework regulates the storage and processing of the data from providers and users in a trustworthy way within the related regulatory frameworks (both legally and ethically). This paper proposes a framework of big data governance that guides organisations to make better data-informed business decisions within the related regularity framework, with close attention paid to data security, privacy, and accessibility. In order to demonstrate this process, the work also presents an example implementation of the framework based on the case study of big data governance in cybersecurity. This framework has the potential to guide the management of big data in different organisations for information sharing and cooperative decision-making.

Enhanced gradient-based local feature descriptors by saliency map for egocentric action recognition

Z Zuo, B Wei, F Chao, Y Qu, Y Peng, L Yang
Journal Applied System Innovation, vol. 2, no. 1, 2019.

Abstract

Egocentric video analysis is an important tool in healthcare that serves a variety of purposes, such as memory aid systems and physical rehabilitation, and feature extraction is an indispensable process for such analysis. Local feature descriptors have been widely applied due to their simple implementation and reasonable efficiency and performance in applications. This paper proposes an enhanced spatial and temporal local feature descriptor extraction method to boost the performance of action classification. The approach allows local feature descriptors to take advantage of saliency maps, which provide insights into visual attention. The effectiveness of the proposed method was validated and evaluated by a comparative study, whose results demonstrated an improved accuracy of around 2%.

An improved fuzzy brain emotional learning model network controller for humanoid robots

W Fang, F Chao, CM Lin, L Yang, C Shang, C Zhou
Journal Frontiers in neurorobotics, vol. 13, 2019.

Abstract

The brain emotional learning (BEL) system was inspired by the biological amygdala-orbitofrontal model to mimic the high speed of the emotional learning mechanism in the mammalian brain, which has been successfully applied in many real-world applications. Despite of its success, such system often suffers from slow convergence for online humanoid robotic control. This paper presents an improved fuzzy BEL model (iFBEL) neural network by integrating a fuzzy neural network (FNN) to a conventional BEL, in an effort to better support humanoid robots. In particular, the system inputs are passed into a sensory and emotional channels that jointly produce the final outputs of the network. The non-linear approximation ability of the iFBEL is achieved by taking the BEL network as the emotional channel. The proposed iFBEL works with a robust controller in generating the hand and gait motion of a humanoid robot. The updating rules of the iFBEL-based controller are composed of two parts, including a sensory channel followed by the updating rules of the conventional BEL model, and the updating rules of the FNN and the robust controller which are derived from the “Lyapunov” function. The experiments on a three-joint robot manipulator and a six-joint biped robot demonstrated the superiority of the proposed system in reference to a conventional proportional-integral-derivative controller and a fuzzy cerebellar model articulation controller, based on the more accurate and faster control performance of the proposed iFBEL.

Semantics Based Web Ranking Using a Robust Weight Scheme

RV Priya, V Vijayakumar, L Yang
Journal International Journal of Web Portals, vol. 11, no. 1, pp. 47-63, 2019.

Abstract

In this paper, HTML tags and attributes are used to determine different structural position of text in a web page. Tags- attributes based models are used to assign a weight to a text that exist in different structural position of web page. Genetic algorithms (GAs), harmony search (HS), and particle swarm optimization (PSO) algorithms are used to select the informative terms using a novel tags-attributes and term frequency weighting scheme. These informative terms with heuristic weight give emphasis to important terms, qualifying how well they semantically explain a webpage and distinguish them from each other. The proposed approach is developed by customizing Terrier and tested over the Clueweb09B, WT10g, .GOV2 and uncontrolled data collections. The performance of the proposed approach is found to be encouraging against five baseline ranking models. The percentage gain of approach achieved is 75-90%, 70-83% and 43-60% in P@5, P@10 and MAP, respectively.

Curvature-based sparse rule base generation for fuzzy rule interpolation

Y Tan, HPH Shum, F Chao, V Vijayakumar, L Yang
Journal Journal of Intelligent & Fuzzy Systems, vol. 36, no. 5, pp. 4201-4214, 2019.

Abstract

Fuzzy inference systems have been successfully applied to many real-world applications. Traditional fuzzy inference systems are only applicable to problems with dense rule bases covering the entire problem domains, whilst fuzzy rule interpolation (FRI) works with sparse rule bases that do not cover certain inputs. Thanks to its ability to work with a rule base with less number of rules, FRI approaches have been utilised as a means to reduce system complexity for complex fuzzy models. This is implemented by removing the rules that can be approximated by their neighbours. Most of the existing fuzzy rule base generation and simplification approaches only target dense rule bases for traditional fuzzy inference systems. This paper proposes a new sparse fuzzy rule base generation method to support FRI. In particular, this approach uses curvature values to identify important rules that cannot be accurately approximated by their neighbouring ones for initialising a compact rule base. The initialised rule base is then optimised using an optimisation algorithm by fine-tuning the membership functions of the involved fuzzy sets. Experiments with a simulation model and a real-world application demonstrate the working principle and the actual performance of the proposed system, with results comparable to the traditional methods using rule bases with more rules.

Augmented YARA rules fused with fuzzy hashing in ransomware triaging

N Naik, P Jenkins, N Savage, L Yang, K Naik, J Song
Conference 2019 IEEE Symposium Series on Computational Intelligence (SSCI), pp. 625-632, 2019

Abstract

Triaging is an initial stage of malware analysis to assess whether a sample is malware or not and the degree of similarity it holds with known malware. It can be applied to any malware category such as ransomware, which is a type of malware that blocks access to a system or data, usually by encrypting it. It has become the main modus operandi for cybercriminals to extort monies from victims due to the growth of cryptocurrencies. Consequently, it severely affects all types of users whether they be from corporates or ordinary home users. Ransomware can be prevented in several different ways, however, the simple and initial step in prevention is its triaging without execution. Several triaging methods are in use such as fuzzy hashing, import hashing and YARA rules, amongst all, YARA rules are one of the most popular and widely used methods. Nonetheless, its success or failure is dependent on the quality of rules employed for malware triaging. This paper performs ransomware triaging using fuzzy hashing, import hashing and YARA rules and demonstrates how YARA rules can be improved using fuzzy hashing to obtain relatively better triaging results. Subsequently, it proposes the augmented YARA rules fused with fuzzy hashing to obtain improved triaging results and performance efficiency in comparison to all three triaging methods individually. Finally, the paper demonstrates how the use of the fused YARA rules can improve triaging results irrespective of the type of malware.

A study of the necessity of signal categorisation in dendritic cell algorithm

N Elisa, F Chao, L Yang
Conference UK Workshop on Computational Intelligence, pp. 210-222, 2019

Abstract

Dendritic Cell Algorithm (DCA) is a binary classifier in the category of artificial immune systems. During its pre-processing phase, DCA requires features to be mapped into three signal categories including safe signal, pathogenic associated molecular pattern, and danger signal, which is usually referred to as signal categorisation. Conventionally, feature-to-signal mapping is performed either manually or automatically by using dimension reduction or feature selection techniques such as principal component analysis and fuzzy rough set theory. The former has been criticised for its potential over-fitting, whilst the latter may suffer from either the loss of underlying feature meaning or impractical for large and complex datasets. This work therefore investigate the necessity of the signal categorisation process by proposing a DCA without the use of signal categorisation but with generalised context detection functions, where the more complex parameters of these functions are learned using the genetic algorithm. This is followed by a comparative study on twelve well-known datasets; the experimental results show overall better performances in terms of accuracy, sensitivity and specificity compared to the conventional DCAs. This confirms that the signal categorisation phase is not necessary, if the weights of the generalised context detection functions can be optimised.

Curvature-based sparse rule base generation for fuzzy interpolation using menger curvature

Z Zuo, J Li, L Yang
Conference 2019 IEEE Symposium Series on Computational Intelligence (SSCI), pp. 625-632, 2019

Abstract

Fuzzy interpolation improves the applicability of fuzzy inference by allowing the utilisation of sparse rule bases. Curvature-based rule base generation approach has been recently proposed to support fuzzy interpolation. Despite the ability to directly generating sparse rule bases from data, the approach often suffers from the high dimensionality of complex inference problems. In this work, a different curvature calculation approach, i.e., the Menger approach, is employed to the curvature-based rule base generation approach in an effort to address the limitation. The experimental results confirm better efficiency and efficacy of the proposed method in generating rule bases on high-dimensional datasets.

A Robotic Chinese Stroke Generation Model Based on Competitive Swarm Optimizer

Q Li, F Chao, X Gao, L Yang, CM Lin, C Shang, C Zhou
Conference UK Workshop on Computational Intelligence, pp. 92-103, 2019

Abstract

The process of neural network based robotic calligraphy involves a trajectory generation process and a robotic manipulator writing process. The writing process of robotic writing cannot be expressed by mathematical expression; therefore, the conventional gradient back-propagation method cannot be directly used to optimize trajectory generation system. This paper alternatively explores the possibility of using competitive swarm optimizer (CSO) algorithm to optimize the neural network used in the robotic calligraphy system. In this paper, a variational auto-encoder network (VAE) including an encoder and a decoder is used to establish the trajectory generation model. The training of the VAE is divided into two steps. In Step 1, the decoder part of VAE network is trained by using the gradient descent method to extract the features of the input strokes. In the second step, the first encoder is used to obtain the image features directly as the input of the decoder, and the writing sequence of stroke trajectory points is obtained directly by the decoder. CSO is applied to train the decoder of VAE. Then the writing sequence is sent to the robot manipulator for writing. Experiments show that the strokes generated by this method can achieve similar but slightly different strokes from the training samples, so that the stroke writing diversity can be retained by VAE. The results also indicate the potential in autonomous action-state space exploration for other real-world applications.

Signal categorisation for dendritic cell algorithm using ga with partial shuffle mutation

N Elisa, L Yang, F Chao
Conference UK Workshop on Computational Intelligence, pp. 529-540, 2019

Abstract

Dendritic Cell Algorithm (DCA) is a bio-inspired system which was specifically developed for anomaly detection problems. In its preprocessing phase, the conventional DC requires domain or expert knowledge to manually categorise the input features for a given dataset into three signal categories termed as safe signal, pathogenic associated molecular pattern and danger signal. The manual preprocessing phase often over-fits the data to the algorithm, which is undesirable. The principal component analysis (PCA) and fuzzy-rough set theory (FRST) based-DCA techniques have been proposed to overcome the aforementioned limitation by automatically categorising the input features to their convenient signal categories. However, the PCA destroys the underlying meaning behind the initial features presented in the input dataset and generates poor classification performance, whilst FRST-DCA is only practical for very simple datasets. Therefore, this study investigates the employment of Genetic Algorithm based on Partial Shuffle Mutation to automatically categorise the input features into the three signal categories. The experimental results of the proposed approach on eleven benchmark datasets have revealed its superiority over other versions of DCA in terms of accuracy, sensitivity and specificity.

Analysis of Big Data Technology for Health Care Services

DS Sathia Raj, B Rawal, L Yang
Preprint arXiv e-prints, arXiv: 1909.03029, 2019

Abstract

Deep learning and other big data technologies have over time become very powerful and accurate. There are algorithms and models developed that have near human accuracy in their task. In health care, the amount of data available is massive and hence, these technologies have a great scope in health care. This paper reviews a few interesting contributions to the field specifically to medical imaging, genomics and patient health records.

Robotic Chinese Calligraphy with Human Preference

F Chao, J Lyu, R Wu, X Gao, C Zhou, L Yang, CM Lin, C Shang
Conference 2019 IEEE Symposium Series on Computational Intelligence (SSCI), pp. 625-632, 2019

Abstract

Robotic Chinese calligraphy is an attempt to lead robots to learn mankind's culture and knowledge. The current research on robotic calligraphy ignores the usage of human preferences. This has restricted robots to produce writing results reflecting personalized styles. This paper proposes a robotic learning approach that introduces a inverse reinforcement learning algorithm with human preferences into a robotic writing system. Through selections of human users, the robot system learns to write Chinese character strokes according to the user's aesthetic preference. Thus, the paper first uses a generative network adopting from the Generative Adversarial Nets to produce a basic writing ability of Chinese strokes for a robot system. Then, the writing results of the robot are captured by the robot's visual device and then presented to the human users as images. Then, the human users give their preferences as the feedbacks of the images, the approach uses the marked images to train a reward predictive mechanism. In the end, the reward predictive mechanism aids the inverse reinforcement learning algorithm to enable the robot to automatically improve its writing ability of Chinese character strokes. Experimental results show that the proposed framework can successfully allow the robot to write Chinese characters strokes in accordance with the human user's preference. In addition, the robot demonstrates a fast learning speed with a small number of human selections. This gives a very promising solution to the robot's learning of more complex movements.

Towards Deep Learning Based Robot Automatic Choreography System

R Wu, W Peng, C Zhou, F Chao, L Yang, CM Lin, C Shang
Conference International Conference on Intelligent Robotics and Applications, pp. 629-640, 2019

Abstract

It is a challenge task to enable a robot to dance according to different types of music. However, two problems have not been well resolved yet: (1) how to assign a dance to a certain type of music, and (2) how to ensure a dancing robot to keep in balance. To tackle these challenges, a robot automatic choreography system based on the deep learning technology is introduced in this paper. First, two deep learning neural network models are built to convert local and global features of music to corresponding features of dance, respectively. Then, an action graph is built based on the collected dance segments; the main function of the action graph is to generate a complete dance sequence based on the dance features generated by the two deep learning models. Finally, the generated dance sequence is performed by a humanoid robot. The experimental results shows that, according to the input music, the proposed model can successfully generate dance sequences that match the input music; also, the robot can maintain its balance while it is dancing. In addition, compared with the dance sequences in the training dataset, the dance sequences generated by the model has reached the level of artificial choreography in both diversity and innovation. Therefore, this method provides a promising solution for robotic choreography automation and design assistance.

A Study of TSK Inference Approaches for Control Problems

J Li, F Chao, L Yang
Conference International Conference on Intelligent Robotics and Applications, pp. 195-207, 2019

Abstract

Fuzzy inference systems provide a simple yet powerful solution to complex non-linear problems, which have been widely and successfully applied in the control field. The TSK-based fuzzy inference approaches, such as the convention TSK, interval type 2 (IT2) TSK and their extensions TSK+ and IT2 TSK+ approaches, are more convenient to be employed in the control field, as they directly produce crisp outputs. This paper systematically reviews those four TSK-based inference approaches, and evaluates them empirically by applying them to a well-known cart centering control problem. The experimental results confirm the power of TSK+ and IT2 TSK+ approaches in enhancing the inference using either dense or sparse rule bases.

Adaptive Activation Function Generation Through Fuzzy Inference for Grooming Text Categorisation

Z Zuo, J Li, B Wei, L Yang, F Chao, N Naik
Conference 2019 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), 2019

Abstract

The activation function is introduced to determine the output of neural networks by mapping the resulting values of neurons into a specific range. The activation functions often suffer from `gradient vanishing', `non zero-centred function outputs', `exploding gradients', and `dead neurons', which may lead to deterioration in the classification performance. This paper proposes an activation function generation approach using the Takagi-Sugeno-Kang inference in an effort to address such challenges. In addition, the proposed method further optimises the coefficients in the activation function using the genetic algorithm such that the activation function can adapt to different applications. This approach has been applied to a digital forensics application of online grooming detection. The evaluations confirm the superiority of the proposed activation function for online grooming detection using an unbalanced data set.

Big data technology in healthcare: a survey

VJ Saglani, BS Rawal, V Vijayakumar, L Yang
Conference 2019 10th IFIP International Conference on New Technologies, Mobility and Security (NTMS), 2019

Abstract

"In God we trust, all others must bring data," quoted by W. Edwards Deming an American engineer, statistician, and professor, have seemed to have taken quite the literal sense in this coming of age world were Robotics, AI, and Machine Learning have become the touchstone of every new burgeoning technology. Today data is not endemic to its source or any organization or an individual; data is omnipresent. Anywhere and everywhere, data flow can be observed, and these data can be saved and analyzed to obtain some confounding observations. To handle the processing of such massive and continually evolving datasets/databases, the concept of Big Data was introduced. This paper discusses how Big Data technologies have been benign in the Healthcare domain.

An intelligent online grooming detection system using AI technologies

P Anderson, Z Zuo, L Yang, Y Qu
Conference 2019 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), 2019

Abstract

The rapid expansion of the Internet has experienced a significant increase in cases of child abuse, as more and more young children have greater access to the Internet. In particular, adults and minors are able to exchange sexually explicit messages and media via a variety of online platforms that are widely available, which leads to an increasing concern of child grooming. Traditionally, the identification of child grooming relies on the analysis and localisation of conversation texts, but this is usually time-consuming and associated with other implications such as psychological pressure on the investigators. Therefore, automatic methods to detect grooming conversations have attracted the attention of many researchers. This paper proposes such a system to identify child grooming in online chat conversations, where the training data of the system were harvested from publicly available information. The data processing is based on a group of AI technologies, including fuzzy-rough feature selection and fuzzy twin support vector machine. Evaluation shows the promise of the proposed approach in identifying online grooming conversations to be implemented in the future after further development to support real-world cases.

Cyberthreat Hunting-Part 1: triaging ransomware using fuzzy hashing, import hashing and YARA rules

N Naik, P Jenkins, N Savage, L Yang
Conference 2019 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), 2019

Abstract

Ransomware is currently one of the most significant cyberthreats to both national infrastructure and the individual, often requiring severe treatment as an antidote. Triaging ran-somware based on its similarity with well-known ransomware samples is an imperative preliminary step in preventing a ransomware pandemic. Selecting the most appropriate triaging method can improve the precision of further static and dynamic analysis in addition to saving significant t ime a nd e ffort. Currently, the most popular and proven triaging methods are fuzzy hashing, import hashing and YARA rules, which can ascertain whether, or to what degree, two ransomware samples are similar to each other. However, the mechanisms of these three methods are quite different and their comparative assessment is difficult. Therefore, this paper presents an evaluation of these three methods for triaging the four most pertinent ransomware categories WannaCry, Locky, Cerber and CryptoWall. It evaluates their triaging performance and run-time system performance, highlighting the limitations of each method.

Cyberthreat hunting-part 2: Tracking ransomware threat actors using fuzzy hashing and fuzzy c-means clustering

N Naik, P Jenkins, N Savage, L Yang
Conference 2019 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), 2019

Abstract

Threat actors are constantly seeking new attack surfaces, with ransomeware being one the most successful attack vectors that have been used for financial gain. T his has been achieved through the dispersion of unlimited polymorphic samples of ransomware whilst those responsible evade detection and hide their identity. Nonetheless, every ransomware threat actor adopts some similar style or uses some common patterns in their malicious code writing, which can be significant evidence contributing to their identification. he first step in attempting to identify the source of the attack is to cluster a large number of ransomware samples based on very little or no information about the samples, accordingly, their traits and signatures can be analysed and identified. T herefore, this paper proposes an efficient fuzzy analysis approach to cluster ransomware samples based on the combination of two fuzzy techniques fuzzy hashing and fuzzy c-means (FCM) clustering. Unlike other clustering techniques, FCM can directly utilise similarity scores generated by a fuzzy hashing method and cluster them into similar groups without requiring additional transformational steps to obtain distance among objects for clustering. Thus, it reduces the computational overheads by utilising fuzzy similarity scores obtained at the time of initial triaging of whether the sample is known or unknown ransomware. The performance of the proposed fuzzy method is compared against k-means clustering and the two fuzzy hashing methods SSDEEP and SDHASH which are evaluated based on their FCM clustering results to understand how the similarity score affects the clustering results.

Adaptive activation function generation for artificial neural networks through fuzzy inference with application in grooming text categorisation

Z Zuo, J Li, B Wei, L Yang, F Chao, N Naik
Conference 2019 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), 2019

Abstract

The activation function is introduced to determine the output of neural networks by mapping the resulting values of neurons into a specific range. The activation functions often suffer from `gradient vanishing', `non zero-centred function outputs', `exploding gradients', and `dead neurons', which may lead to deterioration in the classification performance. This paper proposes an activation function generation approach using the Takagi-Sugeno-Kang inference in an effort to address such challenges. In addition, the proposed method further optimises the coefficients in the activation function using the genetic algorithm such that the activation function can adapt to different applications. This approach has been applied to a digital forensics application of online grooming detection. The evaluations confirm the superiority of the proposed activation function for online grooming detection using an unbalanced data set.

Dendritic cell algorithm enhancement using fuzzy inference system for network intrusion detection

N Elisa, L Yang, X Fu, N Naik
Conference 2019 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), 2019

Abstract

Dendritic cell algorithm (DCA) is an immune-inspired classification algorithm which is developed for the purpose of anomaly detection in computer networks. The DCA uses a weighted function in its context detection phase to process three categories of input signals including safe, danger and pathogenic associated molecular pattern to three output context values termed as co-stimulatory, mature and semi-mature, which are then used to perform classification. The weighted function used by the DCA requires either manually pre-defined weights usually provided by the immunologists, or empirically derived weights from the training dataset. Neither of these is sufficiently flexible to work with different datasets to produce optimum classification result. To address such limitation, this work proposes an approach for computing the three output context values of the DCA by employing the recently proposed TSK+ fuzzy inference system, such that the weights are always optimal for the provided data set regarding a specific application. The proposed approach was validated and evaluated by applying it to the two popular datasets KDD99 and UNSW NB15. The results from the experiments demonstrate that, the proposed approach outperforms the conventional DCA in terms of classification accuracy.

A general transductive regularizer for zero-shot learning

H Mao, H Zhang, Y Long, S Wang, L Yang
Conference 30th British Machine Vision Conference (BMVC) 2019

Abstract

The copyright of this document resides with its authors. Zero Shot Learning (ZSL) has attracted much attention due to its ability to recognize objects of unseen classes, which is realized by transferring knowledge from seen classes through semantic embeddings. Since the seen classes and unseen classes usually have different distributions, conventional inductive ZSL often suffers from the domain shift problem. Transductive ZSL is a type of method for solving such a problem. However, the regularizers of conventional transductive methods are different from each other, and cannot be applied to other methods. In this paper, we propose a General Transductive Regularizer (GTR), which assigns each unlabeled sample to a fixed attribute by defining a Kullback-Leibler Divergence (KLD) objective. To this end, GTR can be easily applied to many compatible linear and deep inductive ZSL models. Extensive experiments on both linear and deep methods are conducted on four popular datasets, and the results show that GTR can significantly improve the performance comparing to its original inductive method, and also outperform some state-of-the-art methods, especially the extension on deep model.

Machine Learning Algorithms for Network Intrusion Detection

J Li, Y Qu, F Chao, HPH Shum, ESL Ho, L Yang
Book Chapter AI in Cybersecurity, pp. 151-179, 2019

Abstract

Network intrusion is a growing threat with potentially severe impacts, which can be damaging in multiple ways to network infrastructures and digital/intellectual assets in the cyberspace. The approach most commonly employed to combat network intrusion is the development of attack detection systems via machine learning and data mining techniques. These systems can identify and disconnect malicious network traffic, thereby helping to protect networks. This chapter systematically reviews two groups of common intrusion detection systems using fuzzy logic and artificial neural networks, and evaluates them by utilizing the widely used KDD 99 benchmark dataset. Based on the findings, the key challenges and opportunities in addressing cyberattacks using artificial intelligence techniques are summarized and future work suggested.

Privacy and security aspects of E-government in smart cities

L Yang, N Elisa, N Eliot
Book Chapter Smart cities cybersecurity and privacy, pp. 89-102, 2019

Abstract

E-government is an indispensable part of a Smart City. Information and communication technologies transform the relationship between citizens, businesses, and government departments, which enables the implementation of e-government, making operational processes efficient and speedy. This chapter investigates the current deployment strategies and the technological solutions of e-government in terms of security and privacy in a Smart City environment; it also identifies the challenges of adoption. In addition, this chapter proposes a decentralized framework based upon blockchain and artificial intelligence to provide a secure and privacy-preserving infrastructure. The proposed framework integrates technologies to provide mutual trust between individuals, businesses, and governments, leading to a greater transparency of activity and less operational overhead. The reduction in process overhead results in lower running costs (therefore increasing revenue) and improves the speed of cross-boundary transactions.

Advances in Computational Intelligence Systems

Z Ju, L Yang, C Yang, A Gegov, D Zhou
Book Advances in Computational Intelligence Systems: Contributions Presented at the 19th UK Workshop on Computational Intelligence, September 4-6, 2019, Portsmouth, UK

Abstract

This book highlights the latest research in computational intelligence and its applications. It covers both conventional and trending approaches in individual chapters on Fuzzy Systems, Intelligence in Robotics, Deep Learning Approaches, Optimization and Classification, Detection, Inference and Prediction, Hybrid Methods, Emerging Intelligence, Intelligent Health Care, and Engineering Data-and Model-Driven Applications. All chapters are based on peer-reviewed contributions presented at the 19th Annual UK Workshop on Computational Intelligence, held in Portsmouth, UK, on 4–6 September 2019. The book offers a valuable reference guide for readers with expertise in computational intelligence or who are seeking a comprehensive and timely review of the latest trends in computational intelligence. Special emphasis is placed on novel methods and their use in a wide range of application areas, updating both academics and professionals on the state of the art.

Soft computing techniques and applications for intelligent multimedia systems

V Varadarajan, V Subramaniyaswamy, L Yang, J Abawajy
Editorial Multimedia Tools and Applications, vol. 78, 2019

Abstract

Of 35 papers submitted to this issue, 7 were eventually accepted after a stringent peer-review process.

Intelligent, smart and scalable cyber-physical systems

V Vijayakumar, V Subramaniyaswamy, J Abawajy, L Yang
Editorial Journal of Intelligent & Fuzzy Systems, vol. 36, no. 5, pp. 3935-3943, 2019

Abstract

The integration of human physical processes with the computation has created a new paradigm called Cyber-Physical Systems (CPS). The CPSs control the physical processes by utilizing the computational intelligence to acquire the deep knowledge of the monitored environment. Hence the CPSs are designed to be intelligent to provide highly accurate decisions and appropriate actions promptly. The rapidly growing interconnections between the virtual and physical worlds and the development of new intelligent techniques have created new opportunities for the research for next-generation CPS, that is intelligent cyber-physical systems (iCPS). The iCPS are large-scale software intensive and pervasive systems, which by combining various data sources (both from physical objects and virtual components) and applying intelligence techniques, can efficiently manage real-world processes and offers a broad range of novel applications and services. By equipping physical objects with interfaces to the virtual world, and incorporating intelligent mechanisms to leverage collaboration between these objects, the boundaries between the physical and virtual worlds become blurred. Interactions occurring in the physical world are capable of changing the processinga behavior in the virtual world, in a causal relationship that can be exploited for the constant improvement of processes. Intelligent, self-aware, self-managing and self-configuring pervasive systems can be built to improve quality of process across a variety of application domains, helping to address a number of contemporary social and environmental issues.

Self-organizing brain emotional learning controller network for intelligent control system of mobile robots

Q. Wu, C. Lin, W. Fang, F. Chao, L. Yang, C. Shang, and C. Zhou
Journal IEEE Access, vol. 6, 2018.

Abstract

The trajectory tracking ability of mobile robots suffers from uncertain disturbances. This paper proposes an adaptive control system consisting of a new type of self-organizing neural network controller for mobile robot control. The newly designed neural network contains the key mechanisms of a typical brain emotional learning controller network and a self-organizing radial basis function network. In this system, the input values are delivered to a sensory channel and an emotional channel, and the two channels interact with each other to generate the final outputs of the proposed network. The proposed network possesses the ability of online generation and elimination of fuzzy rules to achieve an optimal neural structure. The parameters of the proposed network are online tunable by the brain emotional learning rules and gradient descent method; in addition, the stability analysis theory is used to guarantee the convergence of the proposed controller. In the experimentation, a simulated mobile robot was applied to verify the feasibility and effectiveness of the proposed control system. The comparative study using the cutting-edge neural network-based control systems confirms that the proposed network is capable of producing better control performances with high computational efficiency.

A New Dynamic Path Planning Approach for Unmanned Aerial Vehicles

C. Huang, Y. Lan, Y. Liu, W. Zhou, H. Pei, L. Yang, Y. Cheng, Y. Hao, Y. Peng
Journal Complexity, 2018.

Abstract

Dynamic path planning is one of the key procedures for unmanned aerial vehicles (UAV) to successfully fulfill the diversified missions. In this paper, we propose a new algorithm for path planning based on ant colony optimization (ACO) and artificial potential field. In the proposed algorithm, both dynamic threats and static obstacles are taken into account to generate an artificial field representing the environment for collision free path planning. To enhance the path searching efficiency, a coordinate transformation is applied to move the origin of the map to the starting point of the path and in line with the source-destination direction. Cost functions are established to represent the dynamically changing threats, and the cost value is considered as a scalar value of mobile threats which are vectors actually. In the process of searching for an optimal moving direction for UAV, the cost values of path, mobile threats, and total cost are optimized using ant optimization algorithm. The experimental results demonstrated the performance of the new proposed algorithm, which showed that a smoother planning path with the lowest cost for UAVs can be obtained through our algorithm.

Use of human gestures for controlling a mobile robot via adaptive CMAC network and fuzzy logic controller

D. Zhou, M. Shi, F. Chao, C.-M. Lin, L. Yang, C. Shang, and C. Zhou
Journal Neurocomputing, vol. 282, pp. 218-231, 2018.

Abstract

Mobile robots with manipulators have been more and more commonly applied in extreme and hostile environments to assist or even replace human operators for complex tasks. In addition to autonomous abilities, mobile robots need to facilitate the human–robot interaction control mode that enables human users to easily control or collaborate with robots. This paper proposes a system which uses human gestures to control an autonomous mobile robot integrating a manipulator and a video surveillance platform. A human user can control the mobile robot just as one drives an actual vehicle in the vehicle’s driving cab. The proposed system obtains human’s skeleton joints information using a motion sensing input device, which is then recognized and interpreted into a set of control commands. This is implemented, based on the availability of training data set and requirement of in-time performance, by an adaptive cerebellar model articulation controller neural network, a finite state machine, a fuzzy controller and purposely designed gesture recognition and control command generation systems. These algorithms work together implement the steering and velocity control of the mobile robot in real-time. The experimental results demonstrate that the proposed approach is able to conveniently control a mobile robot using virtual driving method, with smooth manoeuvring trajectories in various speeds.

A framework of blockchain-based secure and privacy-preserving E-government system

N Elisa, L Yang, F Chao, Y Cao
Journal Wireless networks, 2018.

Abstract

Electronic government (e-government) uses information and communication technologies to deliver public services to individuals and organisations effectively, efficiently and transparently. E-government is one of the most complex systems which needs to be distributed, secured and privacy-preserved, and the failure of these can be very costly both economically and socially. Most of the existing e-government systems such as websites and electronic identity management systems (eIDs) are centralized at duplicated servers and databases. A centralized management and validation system may suffer from a single point of failure and make the system a target to cyber attacks such as malware, denial of service attacks (DoS), and distributed denial of service attacks (DDoS). The blockchain technology enables the implementation of highly secure and privacy-preserving decentralized systems where transactions are not under the control of any third party organizations. Using the blockchain technology, exiting data and new data are stored in a sealed compartment of blocks (i.e., ledger) distributed across the network in a verifiable and immutable way. Information security and privacy are enhanced by the blockchain technology in which data are encrypted and distributed across the entire network. This paper proposes a framework of a decentralized e-government peer-to-peer (p2p) system using the blockchain technology, which can ensure both information security and privacy while simultaneously increasing the trust of the public sectors. In addition, a prototype of the proposed system is presented, with the support of a theoretical and qualitative analysis of the security and privacy implications of such system.

Gaze-informed egocentric action recognition for memory aid systems

Z. Zuo, L. Yang, Y. Peng, F. Chao, and Y. Qu
JournalIEEE Access, vol. 6, pp. 12894-12904, 2018.

Abstract

Egocentric action recognition has been intensively studied in the fields of computer vision and clinical science with applications in pervasive health-care. The majority of the existing egocentric action recognition techniques utilize the features extracted from either the entire contents or the regions of interest (ROI) in video frames as the inputs of action classifiers. The former might suffer from moving backgrounds or irrelevant foregrounds usually associated with egocentric action videos, while the latter may be impaired by the mismatch between the calculated and the ground truth ROI. This paper proposes a new gaze-informed feature extraction approach, by which the features are extracted from the regions around the gaze points and thus representing the genuine ROI from a first person of view. The activity of daily life can then be classified based only on the identified regions using the extracted gaze-informed features. The proposed approach has been further applied to a memory support system for people with poor memory, such as those with amnesia or dementia, and their carers. The experimental results demonstrate the efficacy of the proposed approach in egocentric action recognition, and thus the potential of the memory support tool in health care.

An extended takagi–sugeno–kang inference system (tsk+) with fuzzy interpolation and its rule base generation

J Li, L Yang, Y Qu, G Sexton
JournalSoft Computing, vol. 22, no. 10, pp. 3155-3170, 2018.

Abstract

A rule base covering the entire input domain is required for the conventional Mamdani inference and Takagi–Sugeno–Kang (TSK) inference. Fuzzy interpolation enhances conventional fuzzy rule inference systems by allowing the use of sparse rule bases by which certain inputs are not covered. Given that almost all of the existing fuzzy interpolation approaches were developed to support the Mamdani inference, this paper presents a novel fuzzy interpolation approach that extends the TSK inference. This paper also proposes a data-driven rule base generation method to support the extended TSK inference system. The proposed system enhances the conventional TSK inference in two ways: (1) workable with incomplete or unevenly distributed data sets or incomplete expert knowledge that entails only a sparse rule base and (2) simplifying complex fuzzy inference systems by using more compact rule bases for complex systems without the sacrificing of system performance. The experimentation shows that the proposed system overall outperforms the existing approaches with the utilisation of smaller rule bases.

Enhanced robotic hand–eye coordination inspired from human-like behavioral patterns

Xin Fu, XiaoJun Zeng, Di Wang, Di Xu, Longzhi Yang
JournalIEEE Transactions on Cognitive and Developmental Systems, Vol. 10, no. 2, 2018

Abstract

Robotic hand-eye coordination is recognized as an important skill to deal with complex real environments. Conventional robotic hand-eye coordination methods merely transfer stimulus signals from robotic visual space to hand actuator space. This paper introduces a reverse method. Build another channel that transfers stimulus signals from robotic hand space to visual space. Based on the reverse channel, a human-like behavior pattern: “Stop-to-Fixate,” is imparted to the robot, thereby giving the robot an enhanced reaching ability. A visual processing system inspired by the human retina structure is used to compress visual information so as to reduce the robot's learning complexity. In addition, two constructive neural networks establish the two sensory delivery channels. The experimental results demonstrate that the robotic system gradually obtains a reaching ability. In particular, when the robotic hand touches an unseen object, the reverse channel successfully drives the visual system to notice the unseen object.

Towards a Robotic Chinese Calligraphy Writing Framework

L Gan, W Fang, F Chao, C Zhou, L Yang, CM Lin, C Shang
Conference 2018 IEEE International Conference on Robotics and Biomimetics (ROBIO), pp. 493-498, 2018

Abstract

Robot calligraphy is often considered as important topic in the research of traditional robots. However, at present, the research of robot calligraphy mainly focuses on the control algorithm and can only write the existing Chinese characters in database with limited writing style which are input to the robot. This paper introduces a robot character writing system to solve this matter. When a user inputs standard Chinese characters, the robot uses other writing style to write rather than “reproduce” the original shape of the characters. Also, the robot can write the Chinese character following human writing habits. The system is implemented by three modules. The stroke disassembled module disassembles the Chinese characters into a number of strokes. The stroke matching module classifies the strokes, and the robot writing actions are completed by the robot writing module. This system can automatically write Chinese characters that do not exist in the robot database. Three common Chinese characters have been used in the experiment for system evaluation. The experimental result shows the success of the proposed framework that the robot writes all the testing characters in the given Chinese calligraphic style.

Towards deep reinforcement learning based chinese calligraphy robot

R Wu, W Fang, F Chao, X Gao, C Zhou, L Yang, CM Lin, C Shang
Conference 2018 IEEE International Conference on Robotics and Biomimetics (ROBIO), pp. 507-512, 2018

Abstract

Learning how to write Chinese character strokes from the stroke images directly, has a great significance to the inheritance of calligraphy art and to imitate the writing style of Chinese calligraphers. However, most of the existing methods directly applied existing samples with action labels. The performance of these methods is often limited by the quality and number of samples. Thus, these methods cannot be used to learn calligraphy from unlabeled samples. To address this problem, a calligraphy robotic model based on deep reinforcement learning is proposed in this paper, which enables a robotic arm to write fundamental Chinese character strokes from stroke images. In the model, writing task is seen as the process of interaction between the robot and the environment. The robot makes appropriate writing action based on the state information provided by the environment. In order to evaluate the writing action of the robot, a reward function is designed on the model. In addition, the stochastic policy gradient method is used in training on the model. Finally, the model was extensively experimented on a stroke data set. Environmental results demonstrate that the proposed model allows a calligraphy robot to successfully write fundamental Chinese character strokes from stroke images. This model provides a promising solution for reconstructing writing actions from images.

Dendritic cell algorithm with fuzzy inference system for input signal generation

N Elisa, J Li, Z Zuo, L Yang
Conference UK workshop on computational intelligence, pp. 203-214, 2018

Abstract

Dendritic cell algorithm (DCA) is a binary classification system developed by abstracting the biological danger theory and the functioning of human dendritic cells. The DCA takes three signals as inputs, including danger, safe and pathogenic associated molecular pattern (PAMP), which are generated in its pre-processing and initialization phase. In particular, after a feature selection process for a given training data set, each selected attribute is assigned to one of the three input signals. Then, these input signals are calculated as the aggregation of their associated features, usually implemented by a simple average function followed by a normalisation process. If a nonlinear relationship exists between a signal and its corresponding selected attributes, the resulting signal using the average function may negatively affect the classification results of the DCA. This work proposes an approach named TSK-DCA to address such limitation by aggregating the assigned features of a signal linearly or non-linearly depending on their inherit relationship using the TSK+ fuzzy inference system. The proposed approach was evaluated and validated using the popular KDD99 data set, and the experimental results indicate the superiority of the proposed approach compared to its conventional counterpart.

Saliency-Informed Spatio-Temporal Vector of Locally Aggregated Descriptors and Fisher Vector for Visual Action Recognition

Z Zuo, D Organisciak, H Shum, L Yang
Conference 29th British Machine Vision Conference (BMVC), 2018

Abstract

Feature encoding has been extensively studied for the task of visual action recognition (VAR). The recently proposed super vector-based encoding methods, such as the Vector of Locally Aggregated Descriptors (VLAD) and the Fisher Vectors (FV), have significantly improved the recognition performance. Despite of the success, they still struggle with the superfluous information that presents during the training stage, which makes the methods computationally expensive when applied to a large number of extracted features. In order to address such challenge, this paper proposes a SaliencyInformed Spatio-Temporal VLAD (SST-VLAD) approach which selects the extracted features corresponding to small amount of videos in the data set by considering both the spatial and temporal video-wise saliency scores; and the same extension principle has also been applied to the FV approach. The experimental results indicate that the proposed feature encoding schemes consistently outperform the existing ones with significantly lower computational cost.

Honeypots that bite back: A fuzzy technique for identifying and inhibiting fingerprinting attacks on low interaction honeypots

N Naik, P Jenkins, R Cooke, L Yang
Conference 2018 IEEE International Conference on fuzzy systems (FUZZ-IEEE), 2018

Abstract

The development of a robust strategy for network security is reliant upon a combination of in-house expertise and for completeness attack vectors used by attackers. A honeypot is one of the most popular mechanisms used to gather information about attacks and attackers. However, low-interaction honeypots only emulate an operating system and services, and are more prone to a fingerprinting attack, resulting in severe consequences such as revealing the identity of the honeypot and thus ending the usefulness of the honeypot forever, or worse, enabling it to be converted into a bot used to attack others. A number of tools and techniques are available both to fingerprint low-interaction honeypots and to defend against such fingerprinting; however, there is an absence of fingerprinting techniques to identify the characteristics and behaviours that indicate fingerprinting is occurring. Therefore, this paper proposes a fuzzy technique to correlate the attack actions and predict the probability that an attack is a fingerprinting attack on the honeypot. Initially, an experimental assessment of the fingerprinting attack on the low- interaction honeypot is performed, and a fingerprinting detection mechanism is proposed that includes the underlying principles of popular fingerprinting attack tools. This implementation is based on a popular and commercially available low-interaction honeypot for Windows - KFSensor. However, the proposed fuzzy technique is a general technique and can be used with any low-interaction honeypot to aid in the identification of the fingerprinting attack whilst it is occurring; thus protecting the honeypot from the fingerprinting attack and extending its life.

Fuzzy Logic Aided Intelligent Threat Detection in Cisco Adaptive Security Appliance 5500 Series Firewalls

N Naik, P Jenkins, B Kerby, J Sloane, L Yang
Conference 2018 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), 2018

Abstract

Cisco Adaptive Security Appliance (ASA) 5500 Series Firewall is amongst the most popular and technically advanced for securing organisational networks and systems. One of its most valuable features is its threat detection function which is available on every version of the firewall running a software version of 8.0(2) or higher. Threat detection operates at layers 3 and 4 to determine a baseline for network traffic, analysing packet drop statistics and generating threat reports based on traffic patterns. Despite producing a large volume of statistical information relating to several security events, further effort is required to mine and visually report more significant information and conclude the security status of the network. There are several commercial off-the-shelf tools available to undertake this task, however, they are expensive and may require a cloud subscription. Furthermore, if the information transmitted over the network is sensitive or requires confidentiality, the involvement of a third party or a third-party tool may place organisational security at risk. Therefore, this paper presents a fuzzy logic aided intelligent threat detection solution, which is a cost-free, intuitive and comprehensible solution, enhancing and simplifying the threat detection process for all. In particular, it employs a fuzzy reasoning system based on the threat detection statistics, and presents results/threats through a developed dashboard user interface, for ease of understanding for administrators and users. The paper further demonstrates the successful utilisation of a fuzzy reasoning system for selected and prioritised security events in basic threat detection, although it can be extended to encompass more complex situations, such as complete basic threat detection, advanced threat detection, scanning threat detection, and customised feature based threat detection.

Dendritic cell algorithm with optimised parameters using genetic algorithm

N Elisa, L Yang, N Naik
Conference 2018 IEEE Congress on Evolutionary Computation (CEC), 2018

Abstract

Intrusion detection systems are developed with the abilities to discriminate between normal and anomalous traffic behaviours. The core challenge in implementing an intrusion detection systems is to determine and stop anomalous traffic behavior precisely before it causes any adverse effects to the network, information systems, or any other hardware and digital assets which forming or in the cyberspace. Inspired by the biological immune system, Dendritic Cell Algorithm (DCA) is a classification algorithm developed for the purpose of anomaly detection based on the danger theory and the functioning of human immune dendritic cells. In its core operation, DCA uses a weighted sum function to derive the output cumulative values from the input signals. The weights used in this function are either derived empirically from the data or defined by users. Due to this, the algorithm opens the doors for users to specify the weights that may not produce optimal result (often accuracy). This paper proposes a weight optimisation approach implemented using the popular stochastic search tool, genetic algorithm. The approach is validated and evaluated using the KDD99 dataset with promising results generated.

Grooming detection using fuzzy-rough feature selection and text classification

Z Zuo, J Li, P Anderson, L Yang, N Naik
Conference 2018 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), 2018

Abstract

Online child grooming detection has recently attracted intensive research interests from both the machine learning community and digital forensics community due to its great social impact. The existing data-driven approaches usually face the challenges of lack of training data and the uncertainty of classes in terms of the classification or decision boundary. This paper proposes a grooming detection approach in an effort to address such uncertainty based on a data set derived from a publicly available profiling data set. In particular, the approach firstly applies the conventional text feature extraction approach in identifying the most significant words in the data set. This is followed by the application of a fuzzy-rough feature selection approach in reducing the high dimensions of the selected words for fast processing, which at the same time addressing the uncertainty of class boundaries. The experimental results demonstrate the efficiency and efficacy of the proposed approach in detecting child grooming.

Interval type-2 tsk+ fuzzy inference system

J Li, L Yang, X Fu, F Chao, Y Qu
Conference 2018 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), 2018

Abstract

Type-2 fuzzy sets and systems can better handle uncertainties compared to its type-1 counterpart, and the widely applied Mamdani and TSK fuzzy inference approaches have been both extended to support interval type-2 fuzzy sets. Fuzzy interpolation enhances the conventional Mamdani and TKS fuzzy inference systems, which not only enables inferences when inputs are not covered by an incomplete or sparse rule base but also helps in system simplification for very complex problems. This paper extends the recently proposed fuzzy interpolation approach TSK+ to allow the utilization of interval type-2 TSK fuzzy rule bases. One illustrative case based on an example problem from the literature demonstrates the working of the proposed system, and the application on the cart centering problem reveals the power of the proposed system. The experimental investigation confirmed that the proposed approach is able to perform fuzzy inferences using either dense or sparse interval type-2 TSK rule bases with promising results generated.

A revised dendritic cell algorithm using k-means clustering

N Elisa, L Yang, Y Qu, F Chao
Conference 2018 IEEE 20th International Conference on High Performance Computing and Communications; IEEE 16th International Conference on Smart City; IEEE 4th International Conference on Data Science and Systems (HPCC/SmartCity/DSS), 2018

Abstract

The most daunting and challenging task in intrusion detection is to distinguishing between normal and malicious traffics effectively. In order to complete such a task, the biological danger theory has appeared to be one of the most appealing immunological models which has been converted to a computer science algorithm, named as Dendritic Cell Algorithm (DCA). To perform a binary classification, the DCA goes through four phases, preprocessing, detection, context assessment and classification. In particular, the context assessment phase is performed by comparing the signal concentration values between mature (i.e., abnormality) and semi-mature (i.e., normality) contexts. The conventional DCA requires a crisp separation between semi-mature and mature cumulative context values. This can be hard if the difference between the two contexts is marginal, which negatively affects the classification accuracy. In addition, it is technically difficult to quantify the actual meaning of semi-mature and mature in the DCA. This paper proposes an approach that integrates the K-Means clustering algorithm to the DCA to map the DCA cumulative semi-mature and mature context values into semi-mature (normal) and mature (anomaly) clusters in order to improve the classification accuracy. The KDD99 data set was utilized in this work for system validation and evaluation, and the experimental results revealed an improvement in the classification accuracy by the proposed approach.

Patient Assessment Assistant Using Augmented Reality

ESL Ho, K McCay, HPH Shum, L Yang, D Sainsbury, P Hodgkinson
Conference Edutainment, 2018

Abstract

Facial symmetry and averageness are key components in quantifying the perception of beauty. In this study, a prototype Augmented Reality (AR) tool is developed on Android OS, to assist plastic surgeons and patients in objectively assessing facial symmetry when planning reconstructive surgical procedures. Speci cally, the tool overlays 4 types of measurements and guidelines over a live video stream to provide the users with useful information interactively. The measurements are computed from the tracked facial landmarks at run-time.

Generative adversarial nets in robotic Chinese calligraphy

F Chao, J Lv, D Zhou, L Yang, CM Lin, C Shang, C Zhou
Conference 2018 IEEE International Conference on Robotics and Automation (ICRA), pp. 1104-1110, 2018

Abstract

Conventional approaches of robotic writing of Chinese character strokes often suffer from limited font generation methods, and thus the writing results often lack of diversity. This has seriously restricted the high quality writing ability of robots. This paper proposes a generative adversarial nets-based calligraphic robotic framework, which enables a robot to learn writing fundamental Chinese strokes with rich diversity and good originality. In particular, the framework considers the learning process of robotic writing as an adversarial procedure which is implemented by three interactive modules including a stroke generation module, a stroke discriminative module and a training module. Noting that the stroke generative module included in the conventional generative adversarial nets cannot solve the non-differentiable problem, the policy gradient commonly used in reinforcement learning is thus adapted in this work to train the generative module by regarding the outputs from the discriminative module as rewards. Experimental results demonstrate that the proposed framework allows a calligraphic robot to successfully write fundamental Chinese strokes with good quality in various styles. The experiment also suggests the proposed approach can achieve human-level stroke writing quality without the requirement of a performance evaluation system. This approach therefore significantly boosts the robotic autonomous creation ability.

Hierarchical quotient spaces-based feature selection

Q Zhang, Y Qu, A Deng, L Yang
Conference 2018 Tenth International Conference on Advanced Computational Intelligence (ICACI), 2018

Abstract

Granular computing is an effective method to deal with imprecise, fuzzy and incomplete information. Commonly, it consists of three popular models: fuzzy sets, rough sets and quotient space. The main interest of the first two methods is to deal with the problem with uncertainty information and that of the latter is to implement the multi-granularity computing. In particular, a quotient space which has a hierarchical structure will be divided into different granules by equivalence relations. In this paper, such hierarchical quotient space is applied to propose a new feature selection method. Specifically, the feature subset is selected by calculating the dependency in the position region of such hierarchical quotient space. The experimental results demonstrate that the performance of the proposed approach outperforms those attainable by typical feature selection methods, in terms of both the size of reduction and classification accuracy.

Adoption of cloud computing in hotel industry as emerging services

E Vella, L Yang, N Anwar, N Jin
Conference International Conference on Information, pp. 218-228, 2018

Abstract

The hotel industry is experiencing forces of change as a result of data explosion, social media, increased individualized expectations by customers. It is thus appealing to study the cloud computing adoption in the hotel industry to respond such changes. This paper reported an investigation on such topic by identifying the cloud computing services and summarising their benefits and challenges in organization, management and operation. The research findings were comparatively studied in reference to the results appeared in the literature. In addition, recommendations were made for both cloud service providers and hotels in strategic planning, investment, and management of cloud-oriented services.

Towards light-weight annotations: Fuzzy interpolative reasoning for zero-shot image classification

Y Long, Y Tan, D Organisciak, L Yang, L Shao
Conference 29th British Machine Vision Conference (BMVC 2018), 2018

Abstract

Despite the recent popularity of Zero-shot Learning (ZSL) techniques, existing approaches rely on ontological engineering with heavy annotations to supervise the transferable attribute model that can go across seen and unseen classes. Moreover, existing cross-sourcing, expert-based, or data-driven attribute annotations (e.g. Word Embeddings) cannot guarantee sufficient description to the visual features, which leads to significant performance degradation. In order to circumvent the expensive attribute annotations while retaining the reliability, we propose a Fuzzy Interpolative Reasoning (FIR) algorithm that can discover inter-class associations from light-weight Simile annotations based on visual similarities between classes. The inferred representation can better bridge the visual-semantic gap and manifest state-of-the-art experimental results.

Generalised Adaptive Fuzzy Rule Interpolation

Longzhi Yang, Fei Chao, Qiang Shen
JournalIEEE Transactions on Fuzzy Systems, vol. 25, no. 4, pp. 839-853, 2017

Abstract

Fuzzy interpolative reasoning strengthens the power of fuzzy inference by the enhancement of the robustness of fuzzy systems and the reduction of the systems' complexity. However, after a series of interpolations, it is possible that multiple object values for a common variable are inferred, leading to inconsistency in interpolated results. Such inconsistencies may result from defective interpolated rules or incorrect interpolative transformations. This paper presents a novel approach for identification and correction of defective rules in interpolative transformations, thereby removing the inconsistencies. In particular, an assumption-based truth-maintenance system (ATMS) is used to record dependences between interpolations, and the underlying technique that the classical general diagnostic engine (GDE) employs for fault localization is adapted to isolate possible faulty interpolated rules and their associated interpolative transformations. From this, an algorithm is introduced to allow for the modification of the original linear interpolation to become first-order piecewise linear. The approach is applied to a realistic problem, which predicates the diarrheal disease rates in remote villages, to demonstrate the potential of this study.

Posture-based and action-based graphs for boxing skill visualization

Y Shen, H Wang, ESL Ho, L Yang, HPH Shum
JournalComputers & Graphics, no. 69, pp. 104-115, 2017

Abstract

Automatic evaluation of sports skills has been an active research area. However, most of the existing research focuses on low-level features such as movement speed and strength. In this work, we propose a framework for automatic motion analysis and visualization, which allows us to evaluate high-level skills such as the richness of actions, the flexibility of transitions and the unpredictability of action patterns. The core of our framework is the construction and visualization of the posture-based graph that focuses on the standard postures for launching and ending actions, as well as the action-based graph that focuses on the preference of actions and their transition probability. We further propose two numerical indices, the Connectivity Index and the Action Strategy Index, to assess skill level according to the graph. We demonstrate our framework with motions captured from different boxers. Experimental results demonstrate that our system can effectively visualize the strengths and weaknesses of the boxers.

A developmental learning approach of mobile manipulator via playing

R Wu, C Zhou, F Chao, Z Zhu, CM Lin, L Yang
JournalFrontiers in Neurorobotics, vol. 11, 2017

Abstract

Inspired by infant development theories, a robotic developmental model combined with game elements is proposed in this paper. This model does not require the definition of specific developmental goals for the robot, but the developmental goals are implied in the goals of a series of game tasks. The games are characterized into a sequence of game modes based on the complexity of the game tasks from simple to complex, and the task complexity is determined by the applications of developmental constraints. Given a current mode, the robot switches to play in a more complicated game mode when it cannot find any new salient stimuli in the current mode. By doing so, the robot gradually achieves it developmental goals by playing different modes of games. In the experiment, the game was instantiated into a mobile robot with the playing task of picking up toys, and the game is designed with a simple game mode and a complex game mode. A developmental algorithm, “Lift-Constraint, Act and Saturate,” is employed to drive the mobile robot move from the simple mode to the complex one. The experimental results show that the mobile manipulator is able to successfully learn the mobile grasping ability after playing simple and complex games, which is promising in developing robotic abilities to solve complex tasks using games.

A robot calligraphy system: From simple to complex writing by human gestures

F Chao, Y Huang, X Zhang, C Shang, L Yang, C Zhou, H Hu, CM Lin
JournalEngineering Applications of Artificial Intelligence, vol. 59, pp. 1-14, 2017

Abstract

Robotic writing is a very challenging task and involves complicated kinematic control algorithms and image processing work. This paper, alternatively, proposes a robot calligraphy system that firstly applies human arm gestures to establish a font database of Chinese character elementary strokes and English letters, then uses the created database and human gestures to write Chinese characters and English words. A three-dimensional motion sensing input device is deployed to capture the human arm trajectories, which are used to build the font database and to train a classifier ensemble. 26 types of human gesture are used for writing English letters, and 5 types of gesture are used to generate 5 elementary strokes for writing Chinese characters. By using the font database, the robot calligraphy system acquires a basic writing ability to write simple strokes and letters. Then, the robot can develop to write complex Chinese characters and English words by following human body movements. The classifier ensemble, which is used to identify each gesture, is implemented through using feature selection techniques and the harmony search algorithm, thereby achieving better classification performance. The experimental evaluations are carried out to demonstrate the feasibility and performance of the proposed method. By following the motion trajectories of the human right arm, the end-effector of the robot can successfully write the English words or Chinese characters that correspond to the arm trajectories.

Fuzzy interpolation systems and applications

L Yang, Z Zuo, F Chao, Y Qu, S Ramakrishnan
Book Chapter Modern fuzzy control systems and its applications, pp. 49-70, 2017

Abstract

Fuzzy inference systems provide a simple yet effective solution to complex non-linear problems, which have been applied to numerous real-world applications with great success. However, conventional fuzzy inference systems may suffer from either too sparse, too complex or imbalanced rule bases, given that the data may be unevenly distributed in the problem space regardless of its volume. Fuzzy interpolation addresses this. It enables fuzzy inferences with sparse rule bases when the sparse rule base does not cover a given input, and it simplifies very dense rule bases by approximating certain rules with their neighbouring ones. This chapter systematically reviews different types of fuzzy interpolation approaches and their variations, in terms of both the interpolation mechanism (inference engine) and sparse rule base generation. Representative applications of fuzzy interpolation in the field of control are also revisited in this chapter, which not only validate fuzzy interpolation approaches but also demonstrate its efficacy and potential for wider applications.

A fall detection/recognition system and an empirical study of gradient-based feature extraction approaches

R Cameron, Z Zuo, G Sexton, L Yang
ConferenceUK Workshop on Computational Intelligence (UKCI), pp. 276-289, 2017.

Abstract

Physically falling down amongst the elder helpless party is one of the most intractable issues in the era of ageing society, which has attracted intensive attentions in academia ranging from clinical research to computer vision studies. This paper proposes a fall detection/recognition system within the realm of computer vision. The proposed system integrates a group of gradient-based local visual feature extraction approaches, including histogram of oriented gradients (HOG), histogram of motion gradients (HMG), histogram of optical flow (HOF), and motion boundary histograms (MBH). A comparative study of the descriptors with the support of an artificial neural network was conducted based on an in-house captured dataset. The experimental results demonstrated the effectiveness of the proposed system and the power of these descriptors in real-world applications.

A computational evaluation system of Chinese calligraphy via extended possibility-probability distribution method

D Zhou, J Ge, R Wu, F Chao, L Yang, C Zhou
Conference 13th International Conference on Natural Computation, Fuzzy Systems and Knowledge Discovery (ICNC-FSKD), 2017.

Abstract

Fuzzy-rough set theory is proven as an effective tool for feature selection. Whilst promising, many state-of-the-art fuzzy-rough feature selection algorithms are time-consuming when dealing with the datasets which have a large quantity of features. In order to address this issue, a λ-partition differentiation entropy fuzzy-rough feature selection (LDE-FRFS) method is proposed in this paper. Such λ-partition differentiation entropy extends the concept of partition differentiation entropy from rough sets to fuzzy-rough sets on the view of a partition of the information system. In this case, it can efficiently gauge the significance of features. Experimental results demonstrate that, by such λ-partition differentiation entropy-based attribute significance, LDE-FRFS outperforms the competitors in terms of both the size of the reduced datasets and the execute time.

Dynamic QoS solution for enterprise networks using TSK fuzzy interpolation

J Li, L Yang, X Fu, F Chao, Y Qu
Conference 2017 IEEE international conference on fuzzy systems (FUZZ-IEEE), 2017.

Abstract

The Quality of Services (QoS) is the measure of data transmission quality and service availability of a network, aiming to maintain the data, especially delay-sensitive data such as VoIP, to be transmitted over the network with the required quality. Major network device manufacturers have each developed their own smart dynamic QoS solutions, such as AutoQoS supported by Cisco, CoS (Class of Service) by Netgear devices, and QoS Maps on SROS (Secure Router Operating System) provided by HP, to maintain the service level of network traffic. Such smart QoS solutions usually only work for manufacture qualified devices and otherwise only a pre-defined static policy mapping can be applied. This paper presents a dynamic QoS solution based on the differentiated services (DiffServ) approach for enterprise networks, which is able to modify the priority level of a packet in real time by adjusting the value of Differentiated Services Code Point (DSCP) in Internet Protocol (IP) header of network packets. This is implemented by a 0-order TSK fuzzy model with a sparse rule base which is developed by considering the current network delay, application desired priority level and user current priority group. DSCP values are dynamically generated by the TSK fuzzy model and updated in real time. The proposed system has been evaluated in a real network environment with promising results generated.

Intrusion detection system by fuzzy interpolation

L Yang, J Li, G Fehringer, P Barraclough, G Sexton, Y Cao
Conference 2017 IEEE international conference on fuzzy systems (FUZZ-IEEE), 2017.

Abstract

Network intrusion detection systems identify malicious connections and thus help protect networks from attacks. Various data-driven approaches have been used in the development of network intrusion detection systems, which usually lead to either very complex systems or poor generalization ability due to the complexity of this challenge. This paper proposes a data-driven network intrusion detection system using fuzzy interpolation in an effort to address the aforementioned limitations. In particular, the developed system equipped with a sparse rule base not only guarantees the online performance of intrusion detection, but also allows the generation of security alerts from situations which are not directly covered by the existing knowledge base. The proposed system has been applied to a well-known data set for system validation and evaluation with competitive results generated.

Fuzzy-rough feature selection based on λ-partition differentiation entropy

Q Sun, Y Qu, A Deng, L Yang
Conference The 13th International Conference on Natural Computation, Fuzzy Systems and Knowledge Discovery, 2017.

Abstract

Fuzzy-rough set theory is proven as an effective tool for feature selection. Whilst promising, many state-of-the-art fuzzy-rough feature selection algorithms are time-consuming when dealing with the datasets which have a large quantity of features. In order to address this issue, a λ-partition differentiation entropy fuzzy-rough feature selection (LDE-FRFS) method is proposed in this paper. Such λ-partition differentiation entropy extends the concept of partition differentiation entropy from rough sets to fuzzy-rough sets on the view of a partition of the information system. In this case, it can efficiently gauge the significance of features. Experimental results demonstrate that, by such λ-partition differentiation entropy-based attribute significance, LDE-FRFS outperforms the competitors in terms of both the size of the reduced datasets and the execute time.

Associated multi-label fuzzy-rough feature selection

Y Qu, Y Rong, A Deng, L Yang
Conference 2017 Joint 17th World Congress of International Fuzzy Systems Association and 9th International Conference on Soft Computing and Intelligent Systems (IFSA-SCIS), 2017.

Abstract

Ahead of the process of selecting a subset of relevant features, the labels commonly need to be combined into a single one for multi-label feature selection. However the existing label combination methods assume that all labels are independent of each other and consequently suffer from high computation complexity. In this paper, association rules implied in the labels are explored to implement a fuzzy-rough feature selection method for multi-label datasets. Specifically, in order to reduce the scale of label and avoid the label overlapping phenomenon, the association rules between labels make the combination of labels collapse to a set of sub-labels. Then each set of sub-labels is regarded as a unique class during the following course of fuzzy-rough feature selection. Empirical results suggest that the quality of the selected features can be improved by the proposed approach compared to the alternative multi-label feature selection algorithms.

Engaging students for the learning and assessment of the advanced computer graphics module using the latest technologies

Y Liu, L Yang, J Ha, B Lu, P Yuen, Y Zhao, R Song
Conference International Conference on Education and New Development, 2017.

Abstract

The advanced computer graphics has been one of the most basic and landmark modules in the field of computer science. It usually covers such topics as core mathematics, lighting and shading, texture mapping, colour and depth, and advanced modeling. All such topics involve mathematics for object modeling and transformation, and programming for object visualization and interaction. While some students are not as good in either mathematics or programming, it is usually a challenge to teach computer graphics to these students effectively. This is because it is difficult for students to link mathematics and programming with what they used to see in video games and the TV advertisements for example and thus they can easily be put off. In this paper, we investigate how the latest technologies can help alleviate the teaching and learning tasks. Instead of selecting the low level programming languages for demonstration and assignment such as Java, Java 3D, C++, or OpenGL, we selected Three.js, which is one of the latest and freely accessible 3D graphics libraries. It has a unique advantage that it provides a seamless interface between the main stream web browsers and 2D/3D graphics. The developed code can be run on a web browser such as Firefox, Chrome, or Safari for testing, debugging and visualization without code changing. The unique design patterns and objectives of Three.js can be very attractive to third party software houses to develop auxiliary functions, methods and tutorials and to make them freely available for the public. Such a unique property of Three.js and its widely available supporting resources are especially helpful to engage students, inspire their learning and facilitate teaching. To evaluate the effectiveness for using Three.js in teaching computer graphics we have set up an assignment for scene modeling in the last 4 years with focuses on the quality of the simulated scene (50%) and the quality of the assignment report (50%). We have evaluated different assessment forms of the module that we taught in the last four years: in 2013-2014 the module consisted of 20% assignment and 80% exam based on Java 3D; in 2014-2015 the same proportion of assignment/exam but based on WebGL, in 2015-2016 the module was 50-50% of assignment and exam but based on Three.js; and in this year the module is 100% assignment based on Three.js. The effectiveness of the module delivery has been evaluated both qualitatively and quantitatively from five aspects: a) average marks of students, b) moderator report, c) module evaluation questionnaire, d) external examiner’s comments and e) examination board recommendations. The results have shown that Three.js is indeed more successful in engaging students for learning and the 100% assignment assessment enables students to focus more on the design and development. This four year result is really encouraging to us as an educational institute to embrace the latest technologies for the delivery of such challenging modules as computer graphics and machine learning.

TSK inference with sparse rule bases

J Li, Y Qu, HPH Shum, L Yang
Conference UK Workshop on Computational Intelligence (UKCI), Advances in Computational Intelligences Systems, pp. 107-123, 2017.

Abstract

The Mamdani and TSK fuzzy models are fuzzy inference engines which have been most widely applied in real-world problems. Compared to the Mamdani approach, the TSK approach is more convenient when the crisp outputs are required. Common to both approaches, when a given observation does not overlap with any rule antecedent in the rule base (which usually termed as a sparse rule base), no rule can be fired, and thus no result can be generated. Fuzzy rule interpolation was proposed to address such issue. Although a number of important fuzzy rule interpolation approaches have been proposed in the literature, all of them were developed for Mamdani inference approach, which leads to the fuzzy outputs. This paper extends the traditional TSK fuzzy inference approach to allow inferences on sparse TSK fuzzy rule bases with crisp outputs directly generated. This extension firstly calculates the similarity degrees between a given observation and every individual rule in the rule base, such that the similarity degrees between the observation and all rule antecedents are greater than 0 even when they do not overlap. Then the TSK fuzzy model is extended using the generated matching degrees to derive crisp inference results. The experimentation shows the promising of the approach in enhancing the TSK inference engine when the knowledge represented in the rule base is not complete.

Integration of fuzzy CMAC and BELC networks for uncertain nonlinear system control

D Zhou, F Chao, CM Lin, L Yang, M Shi, C Zhou
Conference 2017 IEEE international conference on fuzzy systems (FUZZ-IEEE), 2017.

Abstract

This paper develops a fuzzy adaptive control system consisting of a new type of fuzzy neural network and a robust controller for uncertain nonlinear systems. The new designed neural network contains the key mechanisms of a typical fuzzy CMAC network and a brain emotional learning controller network. First, the input values of the new network are delivered to a receptive field structure that is inspired from the fuzzy CMAC. Then, the values are divided into a sensory and an emotional channels; and the two channels interact with each other to generate the final outputs of the proposed network. The parameters of the proposed network are on-line tuned by the brain emotional learning rules; in addition, stability analysis theory is used to guaranty the proposed controller's convergence. In the experimentation, a “Duffing-Holmes” chaotic system and a simulated mobile robot are applied to verify the effectiveness and feasibility of the proposed control system. By comparing with the performances of other neural network based control systems, we believe our proposed network is capable of producing better control performances of complex uncertain nonlinear systems control.

Towards Sparse Rule Base Generation for Fuzzy Rule Interpolation

Y Tan, J Li, M Wonders, F Chao, H Shum, L Yang
ConferenceProceedings of the 2016 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE 2016), Canada, 2016. (IEEE Computational Intelligence Society Outstanding Student Paper Travel Grant Award, Bset Student Paper Award Nomination)

Abstract

Fuzzy inference systems have been successfully applied to many real-world applications. Traditional fuzzy inference systems are only applicable to problems with dense rule bases by which the entire input domain is fully covered, whilst fuzzy rule interpolation (FRI) is also able to work with sparse rule bases that may not cover certain observations. Thanks to its ability to work with fewer rules, fuzzy rule interpolation approaches have also been utilised to reduce system complexity by removing those rules which can be approximated by their neighbouring ones for complex fuzzy models. A number of important fuzzy rule base generation approaches have been proposed in the literature, but the majority of these only target dense rule bases for traditional fuzzy inference systems. This paper proposes a novel sparse fuzzy rule base generation method to support FRI. The approach first identifies important rules that cannot be accurately approximated by their neighbouring ones to initialise the rule base. Then the raw rule base is optimised by fine-tuning the membership functions of the fuzzy sets. Experimentation is conducted to demonstrate the working principles of the proposed system, with results comparable to those of traditional methods.

Experience-Based Rule Base Generation and Adaptation for Fuzzy Interpolation

J Li, H Shum, X Fu, G Sexton, L Yang
ConferenceProceedings of the 2016 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE 2016), Canada, 2016. (IEEE Computational Intelligence Society Outstanding Student Paper Travel Grant Award)

Abstract

Fuzzy modelling has been widely and successfully applied to control problems. Traditional fuzzy modelling requires either complete experts' knowledge or large data sets to generate rule bases such that the input spaces can be fully covered. Although fuzzy rule interpolation (FRI) relaxes this requirement by approximating rules using their neighbouring ones, it is still difficult for some real world applications to obtain sufficient experts' knowledge and/or data to generate a reasonable sparse rule base to support FRI. Also, the generated rule bases are usually fixed and therefore cannot support dynamic situations. In order to address these limitations, this paper presents a novel rule base generation and adaptation system to allow the creation of rule bases with minimal a priori knowledge. This is implemented by adding accurate interpolated rules into the rule base guided by a performance index from the feedback mechanism, also considering the rule's previous experience information as a weight factor in the process of rule selection for FRI. In particular, the selection of rules for interpolation in this work is based on a combined metric of the weight factors and the distances between the rules and a given observation, rather than being simply based on the distances. Two digitally simulated scenarios are employed to demonstrate the working of the proposed system, with promising results generated for both rule base generation and adaptation.

A new fuzzy-rough feature selection algorithm for mammographic risk analysis

Q Guo, Y Qu, A Deng, L Yang
ConferenceProceedings of the 2016 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE 2016), Canada, 2016. (IEEE Computational Intelligence Society Outstanding Student Paper Travel Grant Award, Bset Student Paper Award Nomination)

Abstract

Mammographie risk analysis is a useful means for the early diagnosis of breast cancer. There are many efforts have been devoted to improving the performance of the relevant assessment technologies. This paper presents an invasive weed optimization (IWO) based fuzzy-rough feature selection method for mammographic risk assessment. The advantage of IWO is that the offspring individuals are randomly spread around their parents according to a Gaussian distribution during the evolution process. Such Gaussian distribution is designated with a dynamical standard deviation. Therefore, the optimization algorithm can explore a new solution space aggressively. The diversity of the species can be maintained in the early and middle iterations, and the optimal individuals will be found in the final iteration of feature selection. The mechanism of IWO ensures a global optimal solution for the heuristic search. The performance of IWO is compared against the feature selection methods with ant colony optimization (ACO) and particle swarm optimization (PSO). In the last chapter, the experimental results indicate that the use of IWO entails better performance for the problem of mammographic risk analysis according to both dimensionality reduction and classification accuracy.

Depth Sensor-Based Facial and Body Animation Control

Y Shen, J Zhang, L Yang, H Shum
Book Chapter Handbook of Human Motion, pp. 1943-1958, 2016

Abstract

Depth sensors have become one of the most popular means of generating human facial and posture information in the past decade. By coupling a depth camera and computer vision based recognition algorithms, these sensors can detect human facial and body features in real time. Such a breakthrough has fused many new research directions in animation creation and control, which also has opened up new challenges. In this chapter, we explain how depth sensors obtain human facial and body information. We then discuss on the main challenge on depth sensor-based systems, which is the inaccuracy of the obtained data, and explain how the problem is tackled. Finally, we point out the emerging applications in the field, in which human facial and body feature modeling and understanding is a key research problem.

Fuzzy System Approaches to Negotiation Pricing Decision Support

Xin Fu, XiaoJun Zeng, Di Wang, Di Xu, Longzhi Yang
JournalJournal of Intelligent and Fuzzy Systems, vol. 29, no. 2, pp. 685-699, 2015

Abstract

With the emergence of customisation services, business-to-business price negotiation plays an increasingly important role in economic and management science. Negotiation pricing aims to provide different customers with products/services that perfectly meet their requirements, with the "right" price. In general, pricing managers are responsible for identifying the "right" negotiation price with the goal of maintaining good customer relationship, while maximising profits for companies. However, efficiently and effectively determining the "right" negotiation price boundary is not a simple task; it is often complicated, time-consuming and costly to reach a consensus as the task needs to take a wide variety of pricing factors into consideration, ranging from operation costs, customers' needs to negotiation behaviours. This paper proposes a systematic fuzzy system (FS) approach, for the first time, to provide negotiation price boundary by learning from available historical records, with a goal to release the burden of pricing managers. In addition, when the number of involved influencing factors increases, conventional FS approach easily suffers from the curse of dimensionality. To combat this problem, a novel method, simplified FS with single input and single output modules (SFS-SISOM), is also introduced in this paper to handle high-dimensional negotiation pricing problems. The utility and applicability of this research is illustrated by three experimental datasets that vary from both data dimensionality and the number of training records. The experimental results obtained from two approaches have been compared and analysed based on different aspects, including interpretability, accuracy, generality and applicability.

Multi-layer Lattice Model for Real-Time Dynamic Character Deformation

Naoya Iwamoto, Hubert Shum, Longzhi Yang, Shigeo Morishima
JournalComputer Graphics Forum, vol. 34, no. 7, pp. 99-109, 2015.

Abstract

Due to the recent advancement of computer graphics hardware and software algorithms, deformable characters have become more and more popular in real-time applications such as computer games. While there are mature techniques to generate primary deformation from skeletal movement, simulating realistic and stable secondary deformation such as jiggling of fats remains challenging. On one hand, traditional volumetric approaches such as the finite element method require higher computational cost and are infeasible for limited hardware such as game consoles. On the other hand, while shape matching based simulations can produce plausible deformation in real-time, they suffer from a stiffness problem in which particles either show unrealistic deformation due to high gains, or cannot catch up with the body movement. In this paper, we propose a unified multi-layer lattice model to simulate the primary and secondary deformation of skeleton-driven characters. The core idea is to voxelize the input character mesh into multiple anatomical layers including the bone, muscle, fat and skin. Primary deformation is applied on the bone voxels with lattice-based skinning. The movement of these voxels is propagated to other voxel layers using lattice shape matching simulation, creating a natural secondary deformation. Our multi-layer lattice framework can produce simulation quality comparable to those from other volumetric approaches with a significantly smaller computational cost. It is best to be applied in real-time applications such as console games or interactive animation creation.

Intelligent Home Heating Controller Using Fuzzy Rule Interpolation

J Li, L Yang, HPH Shum, G Sexton, T Yao
ConferenceProceedings of the 2015 UK Workshop on Computational Intelligence, 2015.

Abstract

The reduction of domestic energy waste helps in achieving the legal binding target in the UK that CO2 emissions needs to be reduced by at least 34% below base year (1990) levels by 2020. Space heating consumes about 60% of the household energy consumption, and it has been reported by the Household Electricity Survey from GOV.UK, that 23% of residents leave the heating on while going out. To minimise the waste of heating unoccupied homes, a number of sensor-based and programmable controllers for central heating system have been developed, which can successfully switch off the home heating systems when a property is unoccupied. However, these systems cannot successfully effciently preheat the homes before occupants return without manual inputs or leaving the heating on unnecessarily longer than needed, which has limited the wide application of such devices. In order to address this limitation, this paper proposes a smart home heating controller, which enables a home heating system to effciently preheat the home by successfully predicting the users' home time. In particular, residents' home time is calculated by employing fuzzy rule interpolation, supported by users' historic and current location data from portable devices (commonly smart mobile phones). The proposed system has been applied to a real-world case and promising result has been generated.

A Smart Calendar System Using Multiple Search Techniques

Jake Cowton, Longzhi Yang
ConferenceProceedings of the 2015 UK Workshop on Computational Intelligence, UK, 2015.

Abstract

Calendars are essential for professionals working in industry, government, education and many other fields, which play a key role in the planning and scheduling of people’s day-today events. The majority of existing calendars only provide insight and reminders into what is happening during a certain period of time, but do not offer any actual scheduling functionality that can assist users in creating events to be optimal to their preferences. The burden is on the users to work out when their events should happen, and thus it would be very beneficial to develop a tool to organise personal time to be most efficient based on given tasks, preferences, and constraints, particularly for those people who have generally very busy calendars. This paper proposes a smart calendar system capable of optimising the timing of events to address the limitations of the existing calendar systems. It operates in a tiered format using three search algorithms, namely branch and bound, Hungarian and genetic algorithms, to solve different sized problems with different complexity and features, in an effort to generate a balanced solution between time consumption and optimisation satisfaction. Promising results have shown in the experimentation in personal event planning and scheduling.

Integration Strategies for Toxicity Data from an Empirical Perspective

L Yang, D Neagu
ConferenceProceedings of the 14th Annual UK Workshop on Computational Intelligence (UKCI'14), pp. 1-8, UK, 2014.

Abstract

The recent development of information techniques, especially the state-of-the-art “big data” solutions, enables the extracting, gathering, and processing large amount of toxicity information from multiple sources. Facilitated by this technology advance, a framework named integrated testing strategies (ITS) has been proposed in the predictive toxicology domain, in an effort to intelligently jointly use multiple heterogeneous toxicity data records (through data fusion, grouping, interpolation/extrapolation etc.) for toxicity assessment. This will ultimately contribute to accelerating the development cycle of chemical products, reducing animal use, and decreasing development costs. Most of the current study in ITS is based on a group of consensus processes, termed weight of evidence (WoE), which quantitatively integrate all the relevant data instances towards the same endpoint into an integrated decision supported by data quality. Several WoE implementations for the particular case of toxicity data fusion have been presented in the literature, which are collectively studied in this paper. Noting that these uncertainty handling methodologies are usually not simply developed from conventional probability theory due to the unavailability of big datasets, this paper first investigates the mathematical foundations of these approaches. Then, the investigated data integration models are applied to a representative case in the predictive toxicology domain, with the experimental results compared and analysed.

Closed Form Fuzzy Interpolation with Interval Type-2 Fuzzy Sets

L Yang, C Chen, N Jin, X Fu, and Q Shen
ConferenceProceedings of the 2014 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE'14), pp. 2184-2191, China, 2014. (Nominated best paper)

Abstract

Fuzzy rule interpolation enables fuzzy inference with sparse rule bases by interpolating inference results, and may help to reduce system complexity by removing similar (often redundant) neighbouring rules. In particular, the recently proposed closed form fuzzy interpolation offers a unique approach which guarantees convex interpolated results in a closed form. However, the difficulty in defining the required precise-valued membership functions still poses significant restrictions over the applicability of this approach. Such limitations can be alleviated by employing type-2 fuzzy sets as their membership functions are themselves fuzzy. This paper extends the closed form fuzzy rule interpolation using interval type-2 fuzzy sets. In this way, as illustrated in the experiments, closed form fuzzy interpolation is able to deal with uncertainty in data and knowledge with more flexibility.

Closed Form Fuzzy Interpolation

Longzhi Yang, Qiang Shen
JournalFuzzy Sets and Systems. vol. 225, pp.1-22, 2013.

Abstract

Fuzzy interpolation enhances the robustness of fuzzy systems and helps to reduce systems complexity. Although a number of important fuzzy rule interpolation approaches have been proposed in the literature, most of these approaches cannot be expressed in a closed form. This is usually caused by the effort to avoid possible invalid interpolated results. This paper proposes a different fuzzy rule interpolation approach. It not only can be represented in a closed form but also guarantees that the interpolated results are valid fuzzy sets. This approach is based on a direct use of the extension principle which has been widely utilised for the development of a variety of fuzzy systems. The mathematical properties of the proposed approach are analysed by taking the advantage of the closed form representation. This approach has been applied to a practical problem of predicting diarrhoeal disease rates in remote villages. The results demonstrate the potential of the proposed work in enhancing the robustness of fuzzy interpolation.

Towards a Fuzzy Expert System on Toxicological Data Quality Assessment

L. Yang, D. Neagu, M. Cronin, M. Hewitt, S. Enoch, J. Madden, K. Przybylak
JournalMolecular Informatics, Vol. 32, no. 1, pp. 65-78, 2013.

Abstract

Quality assessment (QA) requires high levels of domain-specific experience and knowledge. QA tasks for toxicological data are usually performed by human experts manually, although a number of quality evaluation schemes have been proposed in the literature. For instance, the most widely utilised Klimisch scheme1 defines four data quality categories in order to tag data instances with respect to their qualities; ToxRTool2 is an extension of the Klimisch approach aiming to increase the transparency and harmonisation of the approach. Note that the processes of QA in many other areas have been automatised by employing expert systems. Briefly, an expert system is a computer program that uses a knowledge base built upon human expertise, and an inference engine that mimics the reasoning processes of human experts to infer new statements from incoming data. In particular, expert systems have been extended to deal with the uncertainty of information by representing uncertain information (such as linguistic terms) as fuzzy sets under the framework of fuzzy set theory and performing inferences upon fuzzy sets according to fuzzy arithmetic. This paper presents an experimental fuzzy expert system for toxicological data QA which is developed on the basis of the Klimisch approach and the ToxRTool in an effort to illustrate the power of expert systems to toxicologists, and to examine if fuzzy expert systems are a viable solution for QA of toxicological data. Such direction still faces great difficulties due to the well‐known common challenge of toxicological data QA that “five toxicologists may have six opinions”. In the meantime, this challenge may offer an opportunity for expert systems because the construction and refinement of the knowledge base could be a converging process of different opinions which is of significant importance for regulatory policy making under the regulation of REACH, though a consensus may never be reached. Also, in order to facilitate the implementation of Weight of Evidence approaches and in silico modelling proposed by REACH, there is a higher appeal of numerical quality values than nominal (categorical) ones, where the proposed fuzzy expert system could help. Most importantly, the deriving processes of quality values generated in this way are fully transparent, and thus comprehensible, for final users, which is another vital point for policy making specified in REACH. Case studies have been conducted and this report not only shows the promise of the approach, but also demonstrates the difficulties of the approach and thus indicates areas for future development.

Towards Model Governance in Predictive Toxicology

A. Palczewska, X. Fu, P. Trundle, L. Yang, D. Neagu, M. Ridley, K. Travis
JournalInternational Journal of Information Managment, vol. 33, no. 3, pp. 567-582, 2013.

Abstract

Efficient management of toxicity information as an enterprise asset is increasingly important for the chemical, pharmaceutical, cosmetics and food industries. Many organisations focus on better information organisation and reuse, in an attempt to reduce the costs of testing and manufacturing in the product development phase. Toxicity information is extracted not only from toxicity data but also from predictive models. Accurate and appropriately shared models can bring a number of benefits if we are able to make effective use of existing expertise. Although usage of existing models may provide high-impact insights into the relationships between chemical attributes and specific toxicological effects, they can also be a source of risk for incorrect decisions. Thus, there is a need to provide a framework for efficient model management. To address this gap, this paper introduces a concept of model governance, that is based upon data governance principles. We extend the data governance processes by adding procedures that allow the evaluation of model use and governance for enterprise purposes. The core aspect of model governance is model representation. We propose six rules that form the basis of a model representation schema, called Minimum Information About a QSAR Model Representation (MIAQMR). As a proof-of-concept of our model governance framework we develop a web application called Model and Data Farm (MADFARM), in which models are described by the MIAQMR-ML markup language.

Toxicity Risk Assessment from Heterogeneous Uncertain Data with Possibility Probability Distribution

Longzhi Yang, Daniel Neagu
ConferenceProceedings of the 2013 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE'13), pp. 1-8, India, 2013.

Abstract

Due to the advance of modern computing technology, decisions can be made based on all the existing related data instances scattered across multiple data storages, such that available information has been entirely taken into consideration. Particularly in the predictive toxicology domain, because of the heterogeneity of data sources, multiple data instances with respect to the same endpoint are usually inconsistent, and the quality (or reliability) of the data instances is typically different. Also, the quantity of data instances is often not sufficient to conduct a study using conventional statistics-based methods. This paper presents a novel risk analysis approach for chemical toxicity assessment which considers all the available heterogeneous data instances in the same time, assisted by their quality (or reliability) values. The system is developed on the basis of possibilityprobability distribution, where the uncertainty of the approximated probability values based on traditional statistics methods is represented by possibility. The uncertainty considered herein is led not only by the statistics on limited small number of data instances, but also by the poor quality (or reliability) of data instances. The possibility-probability distribution is automatically computed from available data instances by employing a modified diffused-interior-outer-set model (where the reliability of data is considered) based on nformation diffusion theory. Toxicity value for a given chemical compound is then estimated as the fuzzy expected value based on the resulted possibility-probability distribution. Toxicity risk with respect to regulatory threshold is also introduced, in order to evaluate the probability of which the toxicity may be classified into a certain regulatory range. The proposed approach is applied to a real-world dataset to illustrate the utility and the potential of the approach in risk assessment of chemical toxicity.

Data Quality in Predictive Toxicology

L Yang, D Neagu, A Palczewska, MJ Ridley
PosterThe Third Annual Meeting of EU Initiative: Safety Evaluation Ultimately Replacing Animal Testing Phase 1 (SEURAT-1), 2013.

Abstract

Due to the advance of database technologies, more and more toxicity data can be accommodated with increasingly complex data structures. Prior to the enjoyment of the convenience brought along with the abundance of data, we are facing a great new challenge: that the chemistry and toxicity data, including the metadata which describe data, need effective and efficient governance in order to be truly beneficial to final users. This poster proposes a general framework for data reliability management, to assist in producing high quality databases, which follow gold standards practically applied and/or systematically specified by regulatory bodies, such as FDA, EPA and OECD. In particular, the multiple representations of each chemical compound in question are compared with those of each chemical compound in the so-called gold standard databases (or inventories), and the consistency result is utilised to examine the quality of chemical information; the processes of assay design, experiment execution, and experimental result documenting, are analysed based on their consistency with internationally accepted guidelines, to help to examine the quality of chemical and toxicological information. The assessed quality values can be utilised to support data fusion in the case that multiple duplicate but not identical data instances present. Two such data fusion mechanisms are also introduced in this poster

COSMOS Database and Data Content

C. Yang, P. Alov, K. Arvidson, M. Checheva, M. Cronin, S. Enoch, E. Fioravanzo, D. Hristozov, K. Jacobs, Y. Lan, J. Madden, C Manelfi, T. Magdziarz, A. Mostrag-Szlchtying, D. Neagu, M. Nelms, J. Rathman, A. Richarz, M. Ridley, O. Sacher, C. Schwab, J. Schwöbel, A. Tarkhov, L. Terfloth, I. Tsakovska, V. Vitcheva, A. Worth, L. Yang
PosterSEURAT-1, 2013.

Abstract

Development of a COSMOS DB to support in silico modelling for cosmetics ingredients and related chemicals

J. Rathman, C. Yang, K. Arvidson, M. Cronin, S. Enoch, D. Hristozov, Y. Lan, J. Madden, D. Neagu, A. Richarz, M. Ridley, O. Sacher, C. Schwab, L. Yang
PosterThe Toxicologist-A Supplement to Toxicological Sciences, vol. 132, no. 189, 2013.

Abstract

Optimisation of Classifier Ensemble for Predictive Toxicology Applications

M Makhtar, L Yang, D Neagu, M Ridley
ConferenceProceedings of the 14th International Conference on Computer Modelling and Simulation (UKSim 2012), pp. 236-241, UK, 2012.

Abstract

Ensembles of classifiers proved potential in getting higher accuracy compared to a single classifier. High diversity in an ensemble may improve the performance results significantly. We propose an ensemble approach which has diversity calculated using disagreement measure of classification output. A CRS (Classifier Ranking System) is introduced for the selection of relevant classifiers. We also propose the Optimisation of Classifiers Ensemble Method (OCEM) technique which applies to the ensemble selection. In this paper, we focus on classification models for predictive toxicology applications, for which computational models are required to replace in vivo experiments. The results show that our method performs well in selecting the relevant ensemble model to improve the prediction from a collection of classifiers.

Towards the integration of heterogeneous uncertain data

L Yang and D Neagu
ConferenceProceedings of the 13th IEEE International Conference on Information Reuse and Integration (IRI-2012), pp. 296-302, USA, 2012.

Abstract

Along with the rapid development of data storing and sharing techniques in terms of both hardware and software, multiple data instances scattered across multiple databases may be available to support one single task, and then making choices of data are necessary from time to time. Research has been conducted on quality or reliability evaluation for individual piece of data assisted by domain knowledge to guide the data selecting processes. However, the choice still can be very difficult if the supporting data instances are contradictory or inconsistent. This paper presents a novel data integration approach based on Credibility Measure, which was developed on the basis of Possibility Measure and Necessity Measure under the framework of fuzzy set theory and fuzzy logic. In particular, the approach is able to combine any new piece of data into the existing decision by an effective credibility revision algorithm such that the revised results have taken all the currently available information into consideration. The proposed approach is applied to a decision problem in the predictive toxicology domain to illustrate the potential in improving the effectiveness of data sharing and the robustness of decisions made from the related data sources.

Data Collation and Curation for the COSMOS Database

L. Yang, Y. Lan, K. Arvidson, M. Cronin, S. Enoch, J. Gasteiger, D. Hristozov, J. Madden, D. Neagu, A. Richarz, J. Rathman, M. Ridley, O. Sacher, C. Schwab, A. Tarkhow, C. Yang
PosterSEURAT-1, 2012.

Abstract

Adaptive Fuzzy Interpolation

Longzhi Yang and Qiang Shen
JournalIEEE Transactions on Fuzzy Systems, vol. 19, no. 6, pp. 1107-1126, 2011.

Abstract

Fuzzy interpolative reasoning strengthens the power of fuzzy inference by the enhancement of the robustness of fuzzy systems and the reduction of the systems' complexity. However, after a series of interpolations, it is possible that multiple object values for a common variable are inferred, leading to inconsistency in interpolated results. Such inconsistencies may result from defective interpolated rules or incorrect interpolative transformations. This paper presents a novel approach for identification and correction of defective rules in interpolative transformations, thereby removing the inconsistencies. In particular, an assumption-based truth-maintenance system (ATMS) is used to record dependences between interpolations, and the underlying technique that the classical general diagnostic engine (GDE) employs for fault localization is adapted to isolate possible faulty interpolated rules and their associated interpolative transformations. From this, an algorithm is introduced to allow for the modification of the original linear interpolation to become first-order piecewise linear. The approach is applied to a realistic problem, which predicates the diarrheal disease rates in remote villages, to demonstrate the potential of this study.

Generalisation of Scale and Move Transformation-Based Fuzzy Interpolation

Qiang Shen , Longzhi Yang
JournalJournal of Advanced Computational Intelligence and Intelligent Informatics, vol. 15, No. 3, pp. 288-298, 2011. (Invited paper, Best Paper Nomination)

Abstract

Fuzzy interpolative reasoning has been extensively studied due to its ability to enhance the robustness of fuzzy systems and reduce system complexity. In particular, the scale and move transformation-based approach is able to handle interpolation with multiple antecedent rules involving triangular, complex polygon, Gaussian and bell-shaped fuzzy membership functions [1]. Also, this approach has been extended to deal with interpolation and extrapolation with multiple multi-antecedent rules [2]. However, the generalised extrapolation approach based on multiple rules may not degenerate back to the basic crisp extrapolation based on two rules. Besides, the approximate function of the extended approach may not be continuous. This paper therefore proposes a new approach to generalising the basic fuzzy interpolation technique of [1] in an effort to eliminate these limitations. Examples are given throughout the paper for illustration and comparative purposes. The result shows that the proposed approach avoids the identified problems, providing more reasonable interpolation and extrapolation.

Adaptive Fuzzy Interpolation with Prioritized Component Candidates

Longzhi Yang, Qiang Shen
ConferenceProceedings of the 2011 IEEE International Conference on Fuzzy Systemses (FUZZ-IEEE'11), pp. 428-435, Taiwan, 2011. (IEEE Computational Intelligence Society Outstanding Student Paper Travel Grant Award)

Abstract

Adaptive fuzzy interpolation strengthens the poten- tial of fuzzy interpolative reasoning. It first identifies all possible sets of faulty fuzzy reasoning components, termed the candidates, each of which may have led to all the contradictory interpolations. It then tries to modify one selected candidate in an effort to remove all the contradictions and thus restore interpolative consistency. This approach assumes that all the candidates are equally likely to be the real culprit. However, this may not be the case in real situations as certain identified reasoning components may be more liable to resulting in inconsistencies than others. This paper extends the adaptive approach by prioritizing all the generated candidates. This is achieved by exploiting the certainty degrees of fuzzy reasoning components and hence of derived propositions. From this, the candidate with the highest priority is modified first. This extension helps to quickly spot the real culprit and thus considerably improves the approach in terms of efficiency.

Automatic Estimation of the Number of Segmentation Groups Based on MI

Ziming Zeng, Wenhui Wang, Longzhi Yang and Reyer Zwiggelaar
ConferenceProceeedings of the 5th Iberian Conference on Pattern Recognition and Image Analysis (IbPRIA 2011), pp. 532-539, Spain, 2011.

Abstract

Clustering is important in medical imaging segmentation. The number of segmentation groups is often needed as an initial condition,but is often unknown. We propose a method to estimate the number of segmentation groups based on mutual information, anisotropicdiffusion model and class-adaptive Gauss-Markov random fields. Initially, anisotropic diffusion is used to decrease the imagenoise. Subsequently, the class-adaptive Gauss-Markov modeling and mutual information are used to determine the number of segmentationgroups. This general formulation enables the method to easily adapt to various kinds of medical images and the associated acquisition artifacts. Experiments on simulated, and multi-model data demonstrate the advantages of the method over the current state-of-the-art approaches.

Adaptive Fuzzy Interpolation with Uncertain Observation and Rule Base

Longzhi Yang, Qiang Shen
ConferenceProdeedings of the 2011 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE'11), pp. 471-478, Taiwan, 2011. (Nominated Best Student Paper, IEEE Computational Intelligence Society Outstanding Student Paper Travel Grant Award)

Abstract

Adaptive fuzzy interpolation strengthens the potential of fuzzy interpolative reasoning. It views interpolation procedures as artificially created system components, and identifies all possible sets of faulty components that may each have led to all detected contradictory results. From this, a modification procedure takes place, which tries to modify each of such components, termed candidates, in an effort to remove all the contradictions and thus restore consistency. This approach assumes that the employed interpolation mechanism is the only cause of contradictions, that is all given observations and rules are believed to be true and fixed. However, this may not be the case in certain real situations. It is common in fuzzy systems that each observation or rule is associated with a certainty degree. This paper extends the adaptive approach by taking into consideration both observations and rules also, treating them as diagnosable and modifiable components in addition to interpolation procedures. Accordingly, the modification procedure is extended to cover the cases of modifying observations or rules in a given rule base along with the modification of fuzzy reasoning components. This extension significantly improves the robustness of the existing adaptive approach.

Adaptive Fuzzy Interpolation and Extrapolation with Multiple-Antecedent Rules

Longzhi Yang, Qiang Shen
ConferenceProceedings of the 2010 IEEE International Conference of Fuzzy Systems (FUZZ-IEEE'11), pp. 1-8, Spain, 2010. (IEEE Computational Intelligence Society Outstanding Student Paper Travel Grant Award, Best Student Paper Award)

Abstract

Adaptive fuzzy interpolation strengthens the potential of fuzzy interpolative reasoning owning to its efficient identification and correction of defective interpolated rules during the interpolation process [11]. This approach assumes that: i) two closest adjacent rules which flank the observation or a previously inferred result are always available; ii) only single-antecedent rules are involved. In practice, however, variable values of these rules may lie just on one side of the observation or inferred result. Also, there may be certain rules with multiple antecedents in the rule base. This paper extends the adaptive approach, in order to cover fuzzy extrapolation and to support rule base with multiple-antecedent rules. Adaptive fuzzy interpolation and extrapolation complement each other, which jointly improve the applicability of fuzzy interpolative reasoning, as it significantly reduces the restriction over the given rule base.

Extending Adaptive Interpolation: From Triangular to Trapezoidal

Longzhi Yang and Qiang Shen
ConferenceProceedings of the 9th UK Workshop on Computational Intelligence, pp. 25-30, UK, 2009.

Abstract

Fuzzy interpolative reasoning strengthens the power of fuzzy inference by enhancing the robustness of fuzzy systems and reducing systems complexity. However, during the interpolation process, it is possible that multiple object values for a common variable are inferred which may lead to inconsistency in interpolated results. A novel approach [10] was recently proposed for identification and correction of defective rules in the transformations computed for interpolation, thereby removing the inconsistencies. However, the implementation of this work is limited to rule models involving triangular fuzzy variables. This paper extends the adaptive approach as presented in [10], by introducing trapezoidal variables in the representation and manipulation of fuzzy rule models. This significantly improves the applicability of adaptive fuzzy interpolation reasoning, as many fuzzy systems are modelled with trapezoidal (as well as triangular) variables.

Towards Adaptive Interpolative Reasoning

Longzhi Yang and Qiang Shen
ConferenceProceedings of the 2009 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE'09), pp. 542-549, South Korea, 2009. (IEEE Computational Intelligence Society Outstanding Student Paper Travel Grant Award)

Abstract

Fuzzy interpolative reasoning has been extensively studied due to its ability to enhance the robustness of fuzzy systems and to reduce system complexity. However, during the interpolation process, it is possible that multiple object values for a common variable are inferred which may lead to inconsistency in interpolated results. Such inconsistencies may result from defective interpolated rules or incorrect interpolative transformations. This paper presents a novel approach for identification and correction of defective rules in transformations, thereby removing the inconsistencies. In particular, an assumption-based truth maintenance system (ATMS) is used to record dependencies between reasoning results and interpolated rules, while the underlying technique that the general diagnostic engine (GDE) employs for fault localization is adapted to isolate possible faulty interpolated rules and their associated interpolative transformations. From this, an algorithm is introduced to allow for the modification of the original linear interpolation to become first-order piecewise linear. The approach is applied to a carefully chosen practical problem to illustrate the potential in strengthening the power of interpolative reasoning.

Current Research Assistant

  • image

    Venkata Nalajarla

    2022 - Present

    Venkata develops Artificial Intelligence solutions for smart dental solutions.

  • image

    Rohit Venugopal

    2021 - Present

    Rohit develops data science solutions for occupational wellbeing.

  • image

    Daniel Turner

    2021 - Present

    Dan develops Artificial Intelligence solutions for geo-technical problems for offshore wind farm.

  • image

    Paras Bhattrai

    2020 - Present

    Paras is working on developing Athelete Monitoring System App, a KTP partnership between Northumbria University and SportsAid. He is a full stack developer developing scalable web applications. His experience include working with Indian and German tech startups. You can find see some of his previous works at GitHub and projects at LinkedIn.

Current PhD Students

  • image

    Yumlembam Rahul

    2021-

    This project mainly focuses on machine learning for cybersecurity, funded by Northumbria University.

  • image

    Jinxin Wang

    2020-

    This project focuses on manufacturing process optimisation using AI for furniture industry.

  • image

    Rebeen Hamad

    2019-

    This project mainly focuses on deep neural network simplification for IoT, funded by Northumbria University.

Recent PhD Students and RAs

  • - Dr. Noe Elisa, PhD student
    - Dr. Shan Shan, PhD student
    - Dr. Jie Li, PhD student
    - Dr. Yijun Shen, PhD student
    - Dr. Yao Tan, PhD student
    - Dr. Zequn Li, PhD student
    - Dr. Naoya Iwamoto, Visiting PhD student
    - Dr. Zheming Zuo, PhD student
    - Dr. Mansoureh Pezhman, PhD student
    - Prof. Mokhairi Makhtar, PhD student
    - Yang Long, Visiting PhD student
    - Dr. Jake Cowton, Research Assistant
    - Aminu Abdulmalik, Reserach Assistant
    - Jason Moore, Research Assistant
    - Dr. Xiuluo Liu, Research Fellow
    - Dr. P. Supraja, Research Fellow
  • Please feel free to contact me if any supervisors or PhD/MSc students (with their supervisors' approval) would like to collaborate with me in co-supervising research projects.

Research Collaborators

Prof. X. Zeng

Prof. X. Zeng at Manchester Uni

+ Follow

Dr. L. Zheng

Research Fellow at Xiamen University

+ Follow

Dr. J. Greensmith

Lecturer at Nottingham University

+ Follow

Dr. Anna Palczewska

Lecturer at Leeds Beckett University

+ Follow

Prof. Daniel Neagu

Professor at Bradford University

+ Follow

Dr. Yanpeng Qu

Associate Professor at Dalian Maritime University

+ Follow

Dr. Yijun Shen

PhD student at Norghumbria University

+ Follow

Dr. Naoya Iwamoto

PhD student at Waseda University

+ Follow

Dr. Fei Chao

Associate Professor at Xiamen University

+ Follow

Dr. Xin Fu

Associate Professor at Xiamen University

+ Follow

Dr. Hubert Shum

Senior Lecturer at Northumbria University

+ Follow

Dr. Mark Flanagan

Information Services Manager at Simpson Group

+ Follow

Ms Elaine Taylor-Whilde

CEO of Nine Health CIC

+ Follow

Dr. Noe Elisa

Lecturer at Dodoma University

+ Follow

Reserach Supervision

Current reserach supervision can be found in "Team & Collaboration" section

Highly motivated PhD candidates and visiting scholars are welcomed to join my team under my supervision

Taught-Programme Computing Projects

Multiple full PhD Scholarships from Russel group universities such as Newcastle and York Universities have been won by BSc final year project students through their reserach-rich projects.

Publications and commercialisation opportunities have been generated from supervised final year proejcts.

Internship and jobs upon graduation have been led by industry-informed final year proejcts.

Industry are encouraged to share your live proejcts for BSc and MSc computing projects.

Taught Modules

Student-centered teaching and learning support

Reserach-led state-of-the-art and fresh teaching contents

Industry-informed genuine case studies and assesmsent scenarios

Inculsive education ensuring maximum learning opportunities for everyone

KV7003: AI and Digital Technologies

KF6007: Artificial Intelligence and Robotics

University Internal Role

Director of Education, strategiclly planning, continuously developing/quality-assuring, and operationally managing:

- Programmes in Newcastle

- Programmes in London

- Programmes in Seoul

- Programmes online in partnership with Pearson

Lead Developer of MSc Artificial Intelligence

Programme Leader of BSc/MComp (Hons)) Computer Networks and Cyber Security (2015-1018)

Programme Leader of BSc/MComp (Hons)) Computer and Digital Forensics (2015-2018)

Event Organisation

General co-Chair, UKCI 2023, Birmingham, UK

Programme Chair, ICMLA 2022, NAssau, The Bahamas

Track Chair of the Securit, Privacy, and Blockchain Track, at ICEBE 2022, Bournemouth, UK

Programme Chair, UKCI 2019, Portsmouth, UK

General Chair, ClooudComp 2018, Exeter, UK

Poster Chair, BMVC 2018, Newcastle upon Tyne, UK

Lead Chair of Annual Special Session Fuzzy Logic Systems for Security and Forensics:

- Special Session 27 "Fuzzy Logic Systems for Security and Forensics" at FUZZ-IEEE 2020 in WCCI 2020

- Special Session 04 "Fuzzy Logic Systems for Security and Forensics" at FUZZ-IEEE 2019

- Special Session 04 "Fuzzy Logic Systems for Security and Forensics" at FUZZ-IEEE 2018 in WCCI 2018

- Special Session 04 "Fuzzy Logic Systems for Security and Forensics" at FUZZ-IEEE 2017

Research Associate Recruitment

A short-term or part-tiem reserach associate is under recruitment. The project is funded by Royal Academy of Engineering, which aims to provide a solution that reduces the capital cost, improves efficiency, and prolongs the lifespan of energy storage systems used in all sizes of Electronic Vehicles, realized in two stages. Firstly, thermal-electric modelling of supercapacitors and batteries will be achieved from first principles to fully understand the dynamic responses of the energy storage components. Then, an adaptive real-time fuzzy neural controller, which is dynamically tuned according to measured and forecast load conditions, will be designed and implemented to optimally manage the power split between the supercapacitor and battery. The research results will not only be validated and evaluated in a lab environment but also tested in real-world conditions with the support of industry partners Averge Technologies (Pty) Ltd and Green Scooter (Pty) Ltd in South Africa.

The applicant must have a good comprehension of fuzzy neural networks and skills to implement such models. It is desirable if the applicant also has a good understanding in battery systems for Electronic Vehicles. Please submit your CV and a one pager statement to longzhi.yang@northumbria.ac.uk by 30th June, 2022.

Please contact Professor Longzhi Yang (longzhi.yang@northumbria.ac.uk) if you have any queries.

PhD Student Position

Highly motivated PhD candidates are welcomed to join, and self-funded student can be accepted all year around.

Undergraduate and MSc Internship Position

Internship students and vistings students are welcome to join us for summer projects.

Visiting Scholar

Motivated visiting scholars are particularly welcome, and we will provide support for candidate to apply the fund for visiting.

Post-doc Research Fellow Position

Motivated post-docs candidates are also welcome, self-funded ones will be particularly encouraged to apply, support will be provided to apply for other funding opportunities, such as Newton International fellowships.

At My Office

CIS Building
Department of Computer and Information Science,
Faculty of Engineering and Environment,
Northumbria University,
Newcastle upon Tyne,
NE1 8ST, UK.