Zhibin Quan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Zhibin Quan has a total of 67 co-authors. Here are the top co-authors by number of shared publications:
Wankou Yang
8 shared publications- Wankou Yang
Southeast University
6 shared publications Weili Zeng
6 shared publications- Weili Zeng
College of Civil Aviation, Nanjing University of Aeronautics and Astronautics, Nanjing 211106, China
5 shared publications Zhiqiang Gao
School of Computer Science and Engineering, Southeast University, Nanjing 211189, China
5 shared publicationsSen Yang
3 shared publications
Latest Publications
- 8 October 2026
- Published by Elsevier in Expert Systems with Applications
- Vol. 334
Beyond post-hoc alignment: Text-conditioned state refinement for weakly supervised referring image segmentation
- 8 October 2026
- Published by Elsevier in Expert Systems with Applications
- Vol. 334
Beyond post-hoc alignment: Text-conditioned state refinement for weakly supervised referring image segmentation
- 18 March 2026
- Published by Elsevier in Expert Systems with Applications
- Vol. 319
Cross-modal semantic token alignment via contrastive learning for weakly-supervised referring image segmentation
- 18 March 2026
- Published by Elsevier in Expert Systems with Applications
- Vol. 319
Cross-modal semantic token alignment via contrastive learning for weakly-supervised referring image segmentation
- 17 September 2025
- Published by Elsevier in Applied Soft Computing
- Vol. 185
AOD-YOLO: A Self-Modulating Multi-Scale Feature Aggregation Mechanism for small object detection in airport surface scenes
- 17 September 2025
- Published by Elsevier in Applied Soft Computing
- Vol. 185
AOD-YOLO: A Self-Modulating Multi-Scale Feature Aggregation Mechanism for small object detection in airport surface scenes
- 11 October 2024
- Published by Elsevier in Engineering Applications of Artificial Intelligence
- Vol. 138
Domain knowledge-powered attention for air traffic management hazardous events classification
- 11 October 2024
- Published by Elsevier in Engineering Applications of Artificial Intelligence
- Vol. 138
Domain knowledge-powered attention for air traffic management hazardous events classification
- 9 February 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE/ACM Transactions on Audio, Speech, and Language Processing
- Vol. 32 (23299290) , 1673-1686
The MorPhEMe Machine: An Addressable Neural Memory for Learning Knowledge-Regularized Deep Contextualized Chinese Embedding
Deep contextualized embeddings, as learned by large pre-training models, have proven highly effective in various downstream natural language processing tasks. However, the embedding space in these large models lacks explicit regularization, leading to underfitting and substantial costs during large-scale training on huge corpora. In this paper, we present a novel approach to learning deep contextualized embeddings, introducing linguistic knowledge regularization. Specifically, our proposed model, MorPhEMe (Morphology and Phonology Embedding Memory), features an external addressable memory with two additional addressable memories for storing morphology and phonology knowledge. MorPhEMe can be seamlessly stacked into a deep architecture. Notably different from existing pre-training models, MorPhEMe boasts two distinctive features: (1) compositional encoding and decompositional decoding facilitated by a dynamic addressing mechanism; and (2) explicit memory embedding regularization through cross-layer memory sharing. Theoretical analysis suggests that the inclusion of morphology and phonology enables MorPhEMe to reduce the modeling complexity of natural language sequences. We evaluate MorPhEMe across a diverse set of Chinese natural language processing tasks, including language modeling, word similarity computation, word analogy reasoning, relation extraction, and machine reading comprehension. Experimental results demonstrate that MorPhEMe, in contrast to state-of-the-art models, achieves remarkable improvements with fewer parameters and rapid convergence.
- 9 February 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE/ACM Transactions on Audio, Speech, and Language Processing
- Vol. 32 (23299290) , 1673-1686
The MorPhEMe Machine: An Addressable Neural Memory for Learning Knowledge-Regularized Deep Contextualized Chinese Embedding
Deep contextualized embeddings, as learned by large pre-training models, have proven highly effective in various downstream natural language processing tasks. However, the embedding space in these large models lacks explicit regularization, leading to underfitting and substantial costs during large-scale training on huge corpora. In this paper, we present a novel approach to learning deep contextualized embeddings, introducing linguistic knowledge regularization. Specifically, our proposed model, MorPhEMe (Morphology and Phonology Embedding Memory), features an external addressable memory with two additional addressable memories for storing morphology and phonology knowledge. MorPhEMe can be seamlessly stacked into a deep architecture. Notably different from existing pre-training models, MorPhEMe boasts two distinctive features: (1) compositional encoding and decompositional decoding facilitated by a dynamic addressing mechanism; and (2) explicit memory embedding regularization through cross-layer memory sharing. Theoretical analysis suggests that the inclusion of morphology and phonology enables MorPhEMe to reduce the modeling complexity of natural language sequences. We evaluate MorPhEMe across a diverse set of Chinese natural language processing tasks, including language modeling, word similarity computation, word analogy reasoning, relation extraction, and machine reading comprehension. Experimental results demonstrate that MorPhEMe, in contrast to state-of-the-art models, achieves remarkable improvements with fewer parameters and rapid convergence.
- 19 December 2023
- Published by Elsevier in Pattern Recognition
- Vol. 149
SED: Searching Enhanced Decoder with switchable skip connection for semantic segmentation
- 19 December 2023
- Published by Elsevier in Pattern Recognition
- Vol. 149
SED: Searching Enhanced Decoder with switchable skip connection for semantic segmentation
- 18 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Vehicles
- Vol. 9 (2) , 3478-3492
Adaptation via Proxy: Building Instance-Aware Proxy for Unsupervised Domain Adaptive 3D Object Detection
3D detection task plays a crucial role in the perception system of intelligent vehicles. LiDAR-based 3D detectors perform well on particular autonomous driving benchmarks, but may poorly generalize to other domains. Existing 3D domain adaptive detection methods usually require annotation-related statistics or continuous refinement of pseudo-labels. The former is not always feasible for practical applications, while the latter lacks sufficient accurate supervision. In this work, we propose a novel unsupervised domain adaptive framework, namely A daptation V ia P roxy ( AVP ), that explicitly leverages cross-domain relationships to generate adequate high-quality samples, thus mitigating domain shifts for existing LiDAR-based 3D detectors. Specifically, we first train the detector on source domain with the curriculum example mining (CEM) strategy to enhance its generalization capability. Then, we integrate the profitable instance knowledge from the source domain with the contextual information from the target domain, to construct the instance-aware proxy, which is a data collection with diverse training scenes and stronger supervision. Finally, we fine-tune the pre-trained detector on the proxy data for further optimizing the detector to overcome domain gaps. To build the instance-aware proxy, two components are proposed, i.e. , the multi-view multi-scale aggregation (MMA) method for producing high-quality pseudo-labels, and the hybrid instance augmentation (HIA) technique for integrating the knowledge from source annotations to enhance supervision. Note that AVP is architecture-agnostic thus it can be easily injected with any LiDAR-based 3D detectors. Extensive experiments on Waymo, nuScenes, KITTI and Lyft demonstrate the superiority of the proposed method over the state-of-the-art approaches for different adaptation scenarios.
- 18 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Vehicles
- Vol. 9 (2) , 3478-3492
Adaptation via Proxy: Building Instance-Aware Proxy for Unsupervised Domain Adaptive 3D Object Detection
3D detection task plays a crucial role in the perception system of intelligent vehicles. LiDAR-based 3D detectors perform well on particular autonomous driving benchmarks, but may poorly generalize to other domains. Existing 3D domain adaptive detection methods usually require annotation-related statistics or continuous refinement of pseudo-labels. The former is not always feasible for practical applications, while the latter lacks sufficient accurate supervision. In this work, we propose a novel unsupervised domain adaptive framework, namely A daptation V ia P roxy ( AVP ), that explicitly leverages cross-domain relationships to generate adequate high-quality samples, thus mitigating domain shifts for existing LiDAR-based 3D detectors. Specifically, we first train the detector on source domain with the curriculum example mining (CEM) strategy to enhance its generalization capability. Then, we integrate the profitable instance knowledge from the source domain with the contextual information from the target domain, to construct the instance-aware proxy, which is a data collection with diverse training scenes and stronger supervision. Finally, we fine-tune the pre-trained detector on the proxy data for further optimizing the detector to overcome domain gaps. To build the instance-aware proxy, two components are proposed, i.e. , the multi-view multi-scale aggregation (MMA) method for producing high-quality pseudo-labels, and the hybrid instance augmentation (HIA) technique for integrating the knowledge from source annotations to enhance supervision. Note that AVP is architecture-agnostic thus it can be easily injected with any LiDAR-based 3D detectors. Extensive experiments on Waymo, nuScenes, KITTI and Lyft demonstrate the superiority of the proposed method over the state-of-the-art approaches for different adaptation scenarios.
- 20 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 3867-3873
Attention based Reinforcement Learning for Efficient Communication under Constraint in Multi-Agent Systems
Efficient communication is a key element of team collaboration in the real-world, and in multi-agent reinforcement learning (MARL) as well. When to communicate with whom and what is the main issue in MARL communication. We propose Graph-attention-Actor Attention-Critic (GA3C) on the basis of actor-critic in which we use graph generator consist of a graph attention networks (GAT) Encoder and multi-layer perceptron (MLP) Decoder to construct directed communication graph determining when to communicate with whom, based on which message processor quickly aggregates and processes messages from agents to achieve communication process. During the training process, critic with attention mechanism makes a more reasonable evaluation of the actor's strategy by dynamically selecting the agents that need attention. Empirical evaluations on the predator-prey environment demonstrate that our method has high-performance compared to other baselines, while having better scalability in more complex environments.
- 20 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 3867-3873
Attention based Reinforcement Learning for Efficient Communication under Constraint in Multi-Agent Systems
Efficient communication is a key element of team collaboration in the real-world, and in multi-agent reinforcement learning (MARL) as well. When to communicate with whom and what is the main issue in MARL communication. We propose Graph-attention-Actor Attention-Critic (GA3C) on the basis of actor-critic in which we use graph generator consist of a graph attention networks (GAT) Encoder and multi-layer perceptron (MLP) Decoder to construct directed communication graph determining when to communicate with whom, based on which message processor quickly aggregates and processes messages from agents to achieve communication process. During the training process, critic with attention mechanism makes a more reasonable evaluation of the actor's strategy by dynamically selecting the agents that need attention. Empirical evaluations on the predator-prey environment demonstrate that our method has high-performance compared to other baselines, while having better scalability in more complex environments.
- 30 March 2023
- Published by Elsevier in Engineering Applications of Artificial Intelligence
- Vol. 123
Enhancing motion visual cues for self-supervised video representation learning
- 30 March 2023
- Published by Elsevier in Engineering Applications of Artificial Intelligence
- Vol. 123
Enhancing motion visual cues for self-supervised video representation learning