Article
Computer Science, Artificial Intelligence
Wei Huang, Mingxi Sun, Liehuang Zhu, Sung-Kwun Oh, Witold Pedrycz
Summary: This study proposes a deep FMNN (DFMNN) based on initialization and optimization operation to overcome the limitations of FMNN, including input order and overlap region problems. DFMNN improves performance by simultaneously designing hyperboxes and implementing deep optimization, outperforming several models previously reported in literature on benchmark datasets.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Yanjuan Ma, Jinhai Liu, Fuming Qu, Hongfei Zhu
Summary: The paper introduces a new method for data classification - FMM-NLA, which uses evolved fuzzy min-max neural network to learn and classify new-labeled data. Compared to traditional methods, FMM-NLA can achieve continuous learning and expand the trained network without retraining all the data. Experimental results demonstrate the effectiveness of FMM-NLA in handling new-labeled data and defect recognition.
APPLIED INTELLIGENCE
(2022)
Article
Computer Science, Artificial Intelligence
Thanh Tung Khuat, Bogdan Gabrys
Summary: This paper proposes an extended online learning algorithm for the General fuzzy min-max neural network (GFMMNN) that can handle datasets with both continuous and categorical features. The algorithm uses the change in entropy values of categorical features to determine if a hyperbox can be expanded to include new training instances.
APPLIED SOFT COMPUTING
(2023)
Article
Computer Science, Artificial Intelligence
Yanjuan Ma, Jinhai Liu, Yan Zhao
Summary: This paper proposes an evolved fuzzy min-max neural network for unknown labeled data classification, which can effectively handle and correct the classification of unknown labeled data. Experimental results demonstrate that the model performs well in handling unknown labeled data and is suitable for practical applications.
NEURAL PROCESSING LETTERS
(2021)
Article
Computer Science, Artificial Intelligence
Farhad Pourpanah, Di Wang, Ran Wang, Chee Peng Lim
Summary: The SSL-FMM model proposed in this paper is a two-stage semisupervised learning model based on FMM networks, which generates hyperboxes through unsupervised and supervised learning stages and improves performance by utilizing a neighborhood-labeling mechanism.
APPLIED SOFT COMPUTING
(2021)
Article
Computer Science, Artificial Intelligence
A. Santhos Kumar, Anil Kumar, Varun Bajaj, Girish Kumar Singh
Summary: The Hyperbox classifier is efficiently implemented using a fuzzy min-max neural network. In the training phase, a set of hyperboxes is constructed based on the input patterns which play a vital role in classification. The use of a secondary training set helps in updating and improving the efficiency of hyperboxes, ultimately leading to higher accuracy rates in classification tasks.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Review
Computer Science, Artificial Intelligence
Omer Nedim Kenger, Eren Ozceylan
Summary: With the rapid development of digital information, the amount of digital data in the universe is growing exponentially, leading to the emergence of new machine learning methods. Learning algorithms using hyperboxes, such as the fuzzy min-max neural network (FMNN), have become increasingly popular and advanced. This paper conducts a bibliometric and network analysis of FMNN literature, identifying trends, challenges, and key points that impact the development of knowledge in this domain.
NEURAL COMPUTING & APPLICATIONS
(2023)
Article
Medicine, General & Internal
Madhura Kalbhor, Swati Shinde, Daniela Elena Popescu, D. Jude Hemanth
Summary: Medical image analysis and classification is an important application of computer vision that assists healthcare professionals in disease prediction. This paper proposes a novel hybrid technique that combines deep learning architectures with machine learning classifiers and fuzzy min-max neural network for feature extraction and Pap-smear image classification.
Article
Computer Science, Artificial Intelligence
Yong Li, Richard Gault, T. Martin McGinnity
Summary: The article introduces a probabilistic fuzzy neural algorithm with a recurrent probabilistic generation module (PFNN-R) to enhance the ability of PFNNs to accommodate noisy data. The back-propagation-based mechanism is utilized to shape the distribution of the probabilistic density function of the fuzzy membership. Through simulation results, it is demonstrated that incorporating recurrency advances the ability of PFNNs to model time-series data with high intensity, random noise.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Thanh Tung Khuat, Fang Chen, Bogdan Gabrys
Summary: Motivated by the practical demands, this article proposes a method to construct classifiers using hyperbox fuzzy sets, maintaining high accuracy through granular inferences and reducing data size significantly. The approach is efficient in terms of training time and predictive performance compared to other fuzzy min-max models and common machine learning algorithms.
IEEE TRANSACTIONS ON FUZZY SYSTEMS
(2021)
Article
Computer Science, Theory & Methods
Martin Gavalec, Zuzana Nemcova, Sergei Sergeev
Summary: By utilizing the concept of (K, L)-eigenvector, the study examines the structure of the max-min eigenspace associated with a specific eigenvalue in the max-min algebra, splitting it into various regions based on the order relations between the eigenvalue and the components of x. The resulting theory of (K, L)-eigenvectors, building upon the foundational work of Gondran and Minoux, offers a comprehensive and detailed description of the entire max-min eigenspace.
FUZZY SETS AND SYSTEMS
(2021)
Article
Computer Science, Artificial Intelligence
Anil Kumar, P. S. V. S. Sai Prasad
Summary: This paper introduces an incremental feature subset selection framework based on fuzzy rough sets, using fuzzy min-max neural network as a preprocessor to handle dynamic data without sacrificing classification performance.
INTERNATIONAL JOURNAL OF APPROXIMATE REASONING
(2021)
Article
Computer Science, Artificial Intelligence
Kumar A. Santhos, A. Kumar, V. Bajaj, G. K. Singh
Summary: Hyperbox classifier has made significant contributions to the field of pattern classification due to its efficiency and transparency. This paper proposes four modifications to the fuzzy min-max (FMM) neural network for increasing the classification accuracy rate. Experimental results demonstrate the improved efficiency of the proposed method.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Computer Science, Information Systems
Thanh Tung Khuat, Bogdan Gabrys
Summary: This paper proposes a method to accelerate the training process of general fuzzy min-max neural network by removing hyperboxes that do not satisfy expansion or aggregation conditions, thus reducing training time. Experimental results show a significant decrease in training time for both online and agglomerative learning algorithms using the proposed method.
INFORMATION SCIENCES
(2021)
Article
Computer Science, Artificial Intelligence
Thanh Tung Khuat, Bogdan Gabrys
Summary: GFMM neural network is an efficient neuro-fuzzy system for classification problems, but the current learning algorithms have limitations in handling only numerical valued features. This paper provides potential approaches to adapting GFMM learning algorithms for classification problems with mixed or categorical features, comparing and evaluating three main methods for handling datasets with mixed features.
Article
Computer Science, Artificial Intelligence
Ngoc Duy Nguyen, Thanh Thi Nguyen, Nhat Truong Pham, Hai Nguyen, Dang Tu Nguyen, Thanh Dang Nguyen, Chee Peng Lim, Michael Johnstone, Asim Bhatti, Douglas Creighton, Saeid Nahavandi
Summary: Reinforcement learning has become an effective approach for building intelligent systems, especially with the introduction of deep learning. This paper proposes a comprehensive software architecture that guides the design of deep reinforcement learning architectures and facilitates the development of realistic reinforcement learning applications.
APPLIED INTELLIGENCE
(2023)
Review
Computer Science, Artificial Intelligence
Farhad Pourpanah, Ran Wang, Chee Peng Lim, Xi-Zhao Wang, Danial Yazdani
Summary: The Artificial Fish Swarm Algorithm (AFSA) is a Swarm Intelligence (SI) methodology inspired by the ecological behaviors of fish schooling. It has been widely used for solving real-world optimization problems due to its flexibility, fast convergence, and insensitivity to initial parameter settings. This paper provides a concise review of continuous AFSA and its improvements, hybrid models, and applications. It also discusses parameter modifications, procedure, and sub-functions of AFSA, along with the reasons for enhancements and comparison results with other methods. Furthermore, hybrid, multi-objective, and dynamic AFSA models for continuous optimization problems are analyzed, and future research directions for advancing AFSA-based models are highlighted.
ARTIFICIAL INTELLIGENCE REVIEW
(2023)
Article
Business
Suriyan Jomthanachai, Wai-Peng Wong, Chee-Peng Lim
Summary: Measuring and improving the efficiency of a sustainable supply chain is crucial for strategic decision making in business. This article introduces an alternative Coherent Data Envelopment Analysis (CoDEA) model for evaluating the efficiency of a sustainable supply chain. The proposed model maintains the value of traditional DEA while overcoming some pitfalls in previous models. Case studies demonstrate the flexibility and reasonableness of the CoDEA model in simple and complex supply chain situations. The inclusion of a dummy decision-making unit (DMU) offers additional benefits in small supply chains.
IEEE TRANSACTIONS ON ENGINEERING MANAGEMENT
(2023)
Article
Engineering, Electrical & Electronic
Houshyar Asadi, Tobias Bellmann, Mohammadreza Chalak Qazani, Shady Mohamed, Chee Peng Lim, Saeid Nahavandi
Summary: A novel decoupled MPC-based algorithm is developed to address the issues related to the existing MPC-based algorithms and generate accurate motion cues. The proposed algorithm is validated through simulation study using MATLAB software.
IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY
(2023)
Article
Computer Science, Interdisciplinary Applications
Abhijit Barman, Pijus Kanti De, Ashis Kumar Chakraborty, Chee Peng Lim, Rubi Das
Summary: With increasing global attention to environmental issues, green supply chain management has become particularly significant. This paper proposes a three-layer green supply chain model with a dual-channel structure and analyzes optimal decisions and pricing strategies under government subsidy and no subsidy. The research findings indicate that government subsidy can reduce costs and enhance supply chain profitability.
MATHEMATICS AND COMPUTERS IN SIMULATION
(2023)
Article
Computer Science, Artificial Intelligence
Long H. Nguyen, Nhat Truong Pham, Van Huong Do, Liu Tai Nguyen, Thanh Tin Nguyen, Hai Nguyen, Ngoc Duy Nguyen, Thanh Thi Nguyen, Sy Dzung Nguyen, Asim Bhatti, Chee Peng Lim
Summary: COVID-19 is a global pandemic caused by the highly contagious SARS-CoV-2 virus. Developing an efficient self-testing service for SARS-CoV-2 at home is crucial, especially with the emergence of the Delta variant. This study introduces a two-stage vision-based framework called Fruit-CoV that detects SARS-CoV-2 infections using recorded cough sounds.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Computer Science, Theory & Methods
Lin Li, Feng Zhang, Jiashuai Zhang, Qiang Hua, Chun-Ru Dong, Chee-Peng Lim
Summary: Unsupervised image clustering is a challenging task in computer vision. In this paper, a deep clustering algorithm based on supported nearest neighbors (SNDC) is proposed to solve the inter-class conflictions problem in deep clustering models. By constructing positive pairs with a support set, SNDC learns more generalized features representation with inherent semantic meaning, leading to superior performance compared to state-of-the-art clustering models on multiple benchmark datasets. The experimental results show accuracy improvements of 6.2% and 20.5% on CIFAR-10 and ImageNet-Dogs respectively.
INTERNATIONAL JOURNAL OF FOUNDATIONS OF COMPUTER SCIENCE
(2023)
Article
Automation & Control Systems
Yang Fei, Peng Shi, Chee Peng Lim, Xin Yuan
Summary: This article comprehensively investigates the challenging problem of robust formation control for omnidirectional robots with model uncertainty and actuator saturation. An observer-based formation controller is designed to ensure the semiglobal uniform ultimate boundedness of the system's formation tracking error. Both global stability and practical finite-time stability are achieved through the observer design. The paper also proposes an adaptive compensator to attenuate the state oscillation caused by the reverse effect of saturated and coupled control input. Numerical simulations and hardware-in-the-loop experiments are conducted to illustrate and evaluate the effectiveness and potentials of the proposed techniques for real-world applications.
IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
(2023)
Article
Engineering, Civil
Mohammad Reza Chalak Qazani, Houshyar Asadi, Yutao Chen, Moloud Abdar, Mansour Karkoub, Shady Mohamed, Chee Peng Lim, Saeid Nahavandi
Summary: Nonlinear model predictive control has been used to consider the nonlinear dynamics model in motion cueing algorithms, allowing precise control of the entire algorithm. However, tuning the weighting parameters in the algorithm requires careful optimization. This work calculates the optimal parameters using cascade optimization and human interaction, improving the motion fidelity compared to a single optimizer.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2023)
Article
Computer Science, Information Systems
Mohammed Nasser Al-Andoli, Kok Swee Sim, Shing Chiang Tan, Pey Yun Goh, Chee Peng Lim
Summary: In this paper, an ensemble-based parallel deep learning classifier is proposed for malware detection. By leveraging five deep learning base models and a neural network as a meta model, the proposed method achieves high accuracy rates on different malware datasets. The parallel implementation also significantly enhances the computational speed.
Article
Computer Science, Information Systems
Mohammed Nasser Al-Andoli, Shing Chiang Tan, Kok Swee Sim, Manjeevan Seera, Chee Peng Lim
Summary: This paper proposes a new parallel ensemble model for fault detection and diagnosis (FDD) tasks using hybrid machine learning and deep learning methods. The model achieves efficient computation through parallel processing, and is validated to have high performance and efficiency in FDD of industrial machinery.
Article
Computer Science, Artificial Intelligence
Sam Slade, Li Zhang, Haoqian Huang, Houshyar Asadi, Chee Peng Lim, Yonghong Yu, Dezong Zhao, Hanhe Lin, Rong Gao
Summary: This paper introduces a novel algorithm, neural inference search (NIS), for optimizing hyperparameters in deep learning segmentation models. NIS incorporates three new search behaviors, including maximized standard deviation velocity prediction, local best velocity prediction, and n-dimensional whirlpool search, to improve performance. Compared with state-of-the-art methods and other search algorithms, NIS-optimized models show significant improvements across multiple performance metrics on segmentation datasets.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Farhad Pourpanah, Moloud Abdar, Yuxuan Luo, Xinlei Zhou, Ran Wang, Chee Peng Lim, Xi-Zhao Wang, Q. M. Jonathan Wu
Summary: Generalized zero-shot learning (GZSL) trains a model to classify data samples when some output classes are unknown. Semantic information of seen and unseen classes is used to bridge the gap between them. This review paper provides an overview, discusses categorization and representative methods, benchmark datasets, applications, and research gaps of GZSL.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Review
Ophthalmology
Shadi Farabi Maleki, Milad Yousefi, Sayeh Afshar, Siamak Pedrammehr, Chee Peng Lim, Ali Jafarizadeh, Houshyar Asadi
Summary: Multiple sclerosis (MS) is a complex autoimmune disease characterized by inflammatory processes, demyelination, neurodegeneration, and axonal damage within the central nervous system (CNS). The integration of artificial intelligence (AI), particularly machine learning (ML) and deep learning (DL) algorithms, with optical coherence tomography (OCT) has shown promise in enhancing MS diagnosis, monitoring disease progression, and revolutionizing patient care. AI-driven approaches have remarkable abilities in accurately detecting and classifying MS-related abnormalities, predicting disease progression, and aiding in personalized treatment planning.
SEMINARS IN OPHTHALMOLOGY
(2023)
Article
Automation & Control Systems
Sasikala Subramaniam, Chee Peng Lim, Rakkiyappan Rajan, Prakash Mani
Summary: Neural networks play a significant role in machine learning and deep learning, and understanding their theoretical properties is important for achieving desired performance in practical applications. This study focuses on the fundamental analysis of neuronal activities through differential modeling and incorporates factors such as time delays, exogenous disturbances, and Markovian-jumping parameters that can affect the stable performance of neural models. Additionally, it addresses the synchronization problem of stochastic neural networks using fractional-derivative Brownian motion and an event-triggered control scheme.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2023)
