Chai Quek
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Chai Quek has a total of 82 co-authors. Here are the top co-authors by number of shared publications:
- Ah-Hwee Tan
School of Computing and Information Systems, Singapore Management University, Singapore, Singapore
7 shared publications - Qi Cao
State Key Laboratory of AI Safety, Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China
5 shared publications - Budhitama Subagdja
School of Computing and Information Systems, Singapore Management University, Singapore, Singapore
4 shared publications - Di Wang
Nanyang technological university, Singapore, Singapore
4 shared publications - Kai Keng Ang
Institute for Infocomm Research, Agency for Science, Technology and Research (A*STAR), 1 Fusionopolis Way, #21-01 Connexis (South Tower), Singapore 138632, Singapore
4 shared publications - Arif Ahmed Sekh
UiT The Arctic University of Norway
2 shared publications
Latest Publications
- 6 October 2026
- Published by Elsevier in Information Sciences
- Vol. 762
Hybrid indicators predicted by an evolving fuzzy interpretable recurrent neural network
- 6 October 2026
- Published by Elsevier in Information Sciences
- Vol. 762
Hybrid indicators predicted by an evolving fuzzy interpretable recurrent neural network
- 30 June 2026
- Published by Elsevier in Applied Soft Computing
- Vol. 202
Stock look-ahead prediction and portfolio management using fuzzy transformer architecture and reinforcement learning
- 30 June 2026
- Published by Elsevier in Applied Soft Computing
- Vol. 202
Stock look-ahead prediction and portfolio management using fuzzy transformer architecture and reinforcement learning
- 30 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 29  (9) , 6406-6416
LLM-Enhanced Multi-Teacher Knowledge Distillation for Modality-Incomplete Emotion Recognition in Daily Healthcare
The critical importance of monitoring and recognizing human emotional states in healthcare has led to a surge in proposals for EEG-based multimodal emotion recognition in recent years. However, practical challenges arise in acquiring EEG signals in daily healthcare settings due to stringent data acquisition conditions, resulting in the issue of incomplete modalities. Existing studies have turned to knowledge distillation as a means to mitigate this problem by transferring knowledge from multimodal networks to unimodal ones. However, these methods are constrained by the use of a single teacher model to transfer integrated feature extraction knowledge, particularly concerning spatial and temporal features in EEG data. To address this limitation, we propose a multi-teacher knowledge distillation framework enhanced with a Large Language Model (LLM), aimed at facilitating effective feature learning in the student network by transferring knowledge of extracting integrated features. Specifically, we employ an LLM as the teacher for extracting temporal features and a graph convolutional neural network for extracting spatial features. To further enhance knowledge distillation, we introduce causal masking and a confidence indicator into the LLM to facilitate the transfer of the most discriminative features. Extensive testing on the DEAP and MAHNOB-HCI datasets demonstrates that our model outperforms existing methods in the modality-incomplete scenario. This study underscores the potential application of large models in this field. The code is publicly available at https://github.com/yuzhezhangEEG/LM-KD .
- 30 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 29  (9) , 6406-6416
LLM-Enhanced Multi-Teacher Knowledge Distillation for Modality-Incomplete Emotion Recognition in Daily Healthcare
The critical importance of monitoring and recognizing human emotional states in healthcare has led to a surge in proposals for EEG-based multimodal emotion recognition in recent years. However, practical challenges arise in acquiring EEG signals in daily healthcare settings due to stringent data acquisition conditions, resulting in the issue of incomplete modalities. Existing studies have turned to knowledge distillation as a means to mitigate this problem by transferring knowledge from multimodal networks to unimodal ones. However, these methods are constrained by the use of a single teacher model to transfer integrated feature extraction knowledge, particularly concerning spatial and temporal features in EEG data. To address this limitation, we propose a multi-teacher knowledge distillation framework enhanced with a Large Language Model (LLM), aimed at facilitating effective feature learning in the student network by transferring knowledge of extracting integrated features. Specifically, we employ an LLM as the teacher for extracting temporal features and a graph convolutional neural network for extracting spatial features. To further enhance knowledge distillation, we introduce causal masking and a confidence indicator into the LLM to facilitate the transfer of the most discriminative features. Extensive testing on the DEAP and MAHNOB-HCI datasets demonstrates that our model outperforms existing methods in the modality-incomplete scenario. This study underscores the potential application of large models in this field. The code is publicly available at https://github.com/yuzhezhangEEG/LM-KD .
- Open Access
- 9 February 2024
- Published by Elsevier in Information Sciences
- Vol. 663
FE-RNN: A fuzzy embedded recurrent neural network for improving interpretability of underlying neural network
- 9 February 2024
- Published by Elsevier in Information Sciences
- Vol. 663
Open AccessFE-RNN: A fuzzy embedded recurrent neural network for improving interpretability of underlying neural network
- Open Access
Adaptive Discounting of Training Time Attacks
Among the most insidious attacks on Reinforcement Learning (RL) solutions are training-time attacks (TTAs) that create loopholes and backdoors in the learned behaviour. Not limited to a simple disruption, constructive TTAs (C-TTAs) are now available, where the attacker forces a specific, target behaviour upon a training RL agent (victim). However, even state-of-the-art C-TTAs focus on target behaviours that could be naturally adopted by the victim if not for a particular feature of the environment dynamics, which C-TTAs exploit. In this work, we show that a C-TTA is possible even when the target behaviour is un-adoptable due to both environment dynamics as well as non-optimality with respect to the victim objective(s). To find efficient attacks in this context, we develop a specialised flavour of the DDPG algorithm, which we term gammaDDPG, that learns this stronger version of C-TTA. gammaDDPG dynamically alters the attack policy planning horizon based on the victim's current behaviour. This improves effort distribution throughout the attack timeline and reduces the effect of uncertainty the attacker has about the victim. To demonstrate the features of our method and better relate the results to prior research, we borrow a 3D grid domain from a state-of-the-art C-TTA for our experiments. Code is available at "bit.ly/github-rb-gDDPG".
Open AccessAdaptive Discounting of Training Time Attacks
Among the most insidious attacks on Reinforcement Learning (RL) solutions are training-time attacks (TTAs) that create loopholes and backdoors in the learned behaviour. Not limited to a simple disruption, constructive TTAs (C-TTAs) are now available, where the attacker forces a specific, target behaviour upon a training RL agent (victim). However, even state-of-the-art C-TTAs focus on target behaviours that could be naturally adopted by the victim if not for a particular feature of the environment dynamics, which C-TTAs exploit. In this work, we show that a C-TTA is possible even when the target behaviour is un-adoptable due to both environment dynamics as well as non-optimality with respect to the victim objective(s). To find efficient attacks in this context, we develop a specialised flavour of the DDPG algorithm, which we term gammaDDPG, that learns this stronger version of C-TTA. gammaDDPG dynamically alters the attack policy planning horizon based on the victim's current behaviour. This improves effort distribution throughout the attack timeline and reduces the effect of uncertainty the attacker has about the victim. To demonstrate the features of our method and better relate the results to prior research, we borrow a 3D grid domain from a state-of-the-art C-TTA for our experiments. Code is available at "bit.ly/github-rb-gDDPG".
- Open Access
- 4 January 2024
- Published by Elsevier in Applied Soft Computing
- Vol. 152
Learning and processing framework using Fuzzy Deep Neural Network for trading and portfolio rebalancing
- 4 January 2024
- Published by Elsevier in Applied Soft Computing
- Vol. 152
Open AccessLearning and processing framework using Fuzzy Deep Neural Network for trading and portfolio rebalancing
- Open Access
- 14 December 2023
- Published by Springer Nature in Complex & Intelligent Systems
- Vol. 10  (2) , 2733-2750
Spatial-temporal episodic memory modeling for ADLs: encoding, retrieval, and prediction
Activities of daily living (ADLs) relate to people’s daily self-care activities, which reflect their living habits and lifestyle. A prior study presented a neural network model called STADLART for ADL routine learning. In this paper, we propose a cognitive model named Spatial-Temporal Episodic Memory for ADL (STEM-ADL), which extends STADLART to encode event sequences in the form of distributed episodic memory patterns. Specifically, STEM-ADL encodes each ADL and its associated contextual information as an event pattern and encodes all events in a day as an episode pattern. By explicitly encoding the temporal characteristics of events as activity gradient patterns, STEM-ADL can be suitably employed for activity prediction tasks. In addition, STEM-ADL can predict both the ADL type and starting time of the subsequent event in one shot. A series of experiments are carried out on two real-world ADL data sets: Orange4Home and OrdonezB, to estimate the efficacy of STEM-ADL. The experimental results indicate that STEM-ADL is remarkably robust in event retrieval using incomplete or noisy retrieval cues. Moreover, STEM-ADL outperforms STADLART and other state-of-the-art models in ADL retrieval and subsequent event prediction tasks. STEM-ADL thus offers a vast potential to be deployed in real-life healthcare applications for ADL monitoring and lifestyle recommendation.
- 14 December 2023
- Published by Springer Nature in Complex & Intelligent Systems
- Vol. 10  (2) , 2733-2750
Open AccessSpatial-temporal episodic memory modeling for ADLs: encoding, retrieval, and prediction
Activities of daily living (ADLs) relate to people’s daily self-care activities, which reflect their living habits and lifestyle. A prior study presented a neural network model called STADLART for ADL routine learning. In this paper, we propose a cognitive model named Spatial-Temporal Episodic Memory for ADL (STEM-ADL), which extends STADLART to encode event sequences in the form of distributed episodic memory patterns. Specifically, STEM-ADL encodes each ADL and its associated contextual information as an event pattern and encodes all events in a day as an episode pattern. By explicitly encoding the temporal characteristics of events as activity gradient patterns, STEM-ADL can be suitably employed for activity prediction tasks. In addition, STEM-ADL can predict both the ADL type and starting time of the subsequent event in one shot. A series of experiments are carried out on two real-world ADL data sets: Orange4Home and OrdonezB, to estimate the efficacy of STEM-ADL. The experimental results indicate that STEM-ADL is remarkably robust in event retrieval using incomplete or noisy retrieval cues. Moreover, STEM-ADL outperforms STADLART and other state-of-the-art models in ADL retrieval and subsequent event prediction tasks. STEM-ADL thus offers a vast potential to be deployed in real-life healthcare applications for ADL monitoring and lifestyle recommendation.
- 1 December 2023
- Published by Springer Nature
- p. 255-284
GEMM-SaFIN(FRIE)++: Explainable Artificial Intelligence Visualisation System with Episodic Memory
- 1 December 2023
- Published by Springer Nature
- p. 255-284
GEMM-SaFIN(FRIE)++: Explainable Artificial Intelligence Visualisation System with Episodic Memory
- 7 February 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks and Learning Systems
- Vol. 35  (8) , 10288-10300
Value-Based Subgoal Discovery and Path Planning for Reaching Long-Horizon Goals
Learning to reach long-horizon goals in spatial traversal tasks is a significant challenge for autonomous agents. Recent subgoal graph-based planning methods address this challenge by decomposing a goal into a sequence of shorter-horizon subgoals. These methods, however, use arbitrary heuristics for sampling or discovering subgoals, which may not conform to the cumulative reward distribution. Moreover, they are prone to learning erroneous connections (edges) between subgoals, especially those lying across obstacles. To address these issues, this article proposes a novel subgoal graph-based planning method called learning subgoal graph using value-based subgoal discovery and automatic pruning (LSGVP). The proposed method uses a subgoal discovery heuristic that is based on a cumulative reward (value) measure and yields sparse subgoals, including those lying on the higher cumulative reward paths. Moreover, LSGVP guides the agent to automatically prune the learned subgoal graph to remove the erroneous edges. The combination of these novel features helps the LSGVP agent to achieve higher cumulative positive rewards than other subgoal sampling or discovery heuristics, as well as higher goal-reaching success rates than other state-of-the-art subgoal graph-based planning methods.
- 7 February 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks and Learning Systems
- Vol. 35  (8) , 10288-10300
Value-Based Subgoal Discovery and Path Planning for Reaching Long-Horizon Goals
Learning to reach long-horizon goals in spatial traversal tasks is a significant challenge for autonomous agents. Recent subgoal graph-based planning methods address this challenge by decomposing a goal into a sequence of shorter-horizon subgoals. These methods, however, use arbitrary heuristics for sampling or discovering subgoals, which may not conform to the cumulative reward distribution. Moreover, they are prone to learning erroneous connections (edges) between subgoals, especially those lying across obstacles. To address these issues, this article proposes a novel subgoal graph-based planning method called learning subgoal graph using value-based subgoal discovery and automatic pruning (LSGVP). The proposed method uses a subgoal discovery heuristic that is based on a cumulative reward (value) measure and yields sparse subgoals, including those lying on the higher cumulative reward paths. Moreover, LSGVP guides the agent to automatically prune the learned subgoal graph to remove the erroneous edges. The combination of these novel features helps the LSGVP agent to achieve higher cumulative positive rewards than other subgoal sampling or discovery heuristics, as well as higher goal-reaching success rates than other state-of-the-art subgoal graph-based planning methods.
- 21 September 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Affective Computing
- Vol. 13  (4) , 1827-1839
EEG-Video Emotion-Based Summarization: Learning With EEG Auxiliary Signals
Video summarization is the process of selecting a subset of informative keyframes to expedite storytelling with limited loss of information. In this paper, we propose an EEG-Video Emotion-based Summarization (EVES) model based on a multimodal deep reinforcement learning (DRL) architecture that leverages neural signals to learn visual interestingness to produce quantitatively and qualitatively better video summaries. As such, EVES does not learn from the expensive human annotations but the multimodal signals. Furthermore, to ensure the temporal alignment and minimize the modality gap between the visual and EEG modalities, we introduce a Time Synchronization Module (TSM) that uses an attention mechanism to transform the EEG representations onto the visual representation space. We evaluate the performance of EVES on the TVSum and SumMe datasets. Based on the rank order statistics benchmarks, the experimental results show that EVES outperforms the unsupervised models and narrows the performance gap with supervised models. Furthermore, the human evaluation scores show that EVES receives a higher rating than the state-of-the-art DRL model DR-DSN by 11.4% on the coherency of the content and 7.4% on the emotion-evoking content. Thus, our work demonstrates the potential of EVES in selecting interesting content that is both coherent and emotion-evoking.
- 21 September 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Affective Computing
- Vol. 13  (4) , 1827-1839
EEG-Video Emotion-Based Summarization: Learning With EEG Auxiliary Signals
Video summarization is the process of selecting a subset of informative keyframes to expedite storytelling with limited loss of information. In this paper, we propose an EEG-Video Emotion-based Summarization (EVES) model based on a multimodal deep reinforcement learning (DRL) architecture that leverages neural signals to learn visual interestingness to produce quantitatively and qualitatively better video summaries. As such, EVES does not learn from the expensive human annotations but the multimodal signals. Furthermore, to ensure the temporal alignment and minimize the modality gap between the visual and EEG modalities, we introduce a Time Synchronization Module (TSM) that uses an attention mechanism to transform the EEG representations onto the visual representation space. We evaluate the performance of EVES on the TVSum and SumMe datasets. Based on the rank order statistics benchmarks, the experimental results show that EVES outperforms the unsupervised models and narrows the performance gap with supervised models. Furthermore, the human evaluation scores show that EVES receives a higher rating than the state-of-the-art DRL model DR-DSN by 11.4% on the coherency of the content and 7.4% on the emotion-evoking content. Thus, our work demonstrates the potential of EVES in selecting interesting content that is both coherent and emotion-evoking.