Wentao Tan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Wentao Tan has a total of 36 co-authors. Here are the top co-authors by number of shared publications:
- Lei Zhu
Tongji University, Shanghai, China
4 shared publications - Jingjing Li
University of Electronic Science and Technology of China, Chengdu, China
3 shared publications - Bowen Wang
Tongji University, Shanghai, China
2 shared publications - Fengling Li
Faculty of Engineering and Information Technology, University of Technology Sydney, Sydney, NSW, Australia
2 shared publications - Heng Tao Shen
Tongji University, Shanghai, China
2 shared publications - Yipeng Chen
Tongji University, Shanghai, China
2 shared publications
Latest Publications
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 912-920
Non-Markovian Long-Horizon Robot Manipulation via Keyframe Chaining
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 912-920
Open AccessNon-Markovian Long-Horizon Robot Manipulation via Keyframe Chaining
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 9402-9411
BLM 1 : A Boundless Large Model for Cross-Space, Cross-Task, and Cross-Embodiment Learning
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 9402-9411
Open AccessBLM 1 : A Boundless Large Model for Cross-Space, Cross-Task, and Cross-Embodiment Learning
- Open Access
- 27 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in TechRxiv
Towards Generalist Embodied AI: A Survey on World Models for VLA Agents
Vision-language-action (VLA) models represent a critical milestone toward generalist embodied AI. However, they often struggle to precisely capture physical dynamics, verify plan executability, and overcome data scarcity. To address these limitations, world models are introduced as future predictors that incorporate rich physical priors to effectively guide VLA agents toward physically grounded action generation. While this paradigm is emerging as a pivotal research direction, the field currently lacks a systematic overview. Therefore, we present the first survey specifically dedicated to world models for VLA agents. We propose a unified taxonomy that organizes existing methods into four paradigms: world planner, world action model, world synthesizer, and world simulator, and review their respective evolutionary trajectories. Furthermore, we summarize the ecosystem of foundation models, benchmarks, and evaluation metrics to support future research. Finally, we highlight key challenges and promising directions to facilitate the advancement of world models for generalist VLA agents.
- 27 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in TechRxiv
Open AccessTowards Generalist Embodied AI: A Survey on World Models for VLA Agents
Vision-language-action (VLA) models represent a critical milestone toward generalist embodied AI. However, they often struggle to precisely capture physical dynamics, verify plan executability, and overcome data scarcity. To address these limitations, world models are introduced as future predictors that incorporate rich physical priors to effectively guide VLA agents toward physically grounded action generation. While this paradigm is emerging as a pivotal research direction, the field currently lacks a systematic overview. Therefore, we present the first survey specifically dedicated to world models for VLA agents. We propose a unified taxonomy that organizes existing methods into four paradigms: world planner, world action model, world synthesizer, and world simulator, and review their respective evolutionary trajectories. Furthermore, we summarize the ecosystem of foundation models, benchmarks, and evaluation metrics to support future research. Finally, we highlight key challenges and promising directions to facilitate the advancement of world models for generalist VLA agents.
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 28, 12207-12218
Meta-Semantic Induction for Generalized Cross-Modal Hashing
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 28, 12207-12218
Meta-Semantic Induction for Generalized Cross-Modal Hashing
- 7 November 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 48 (3) , 2954-2969
Dynamic Bit-Wise Semantic Transformer Hashing for Multi-Modal Retrieval
Multi-modal hashing aims to succinctly encode heterogeneous modalities into binary hash codes, facilitating efficient multimedia retrieval characterized by low storage demands and high retrieval speed. Despite the commendable achievements of existing methods, they still face three crucial challenges: 1) Inadequate bridging of the heterogeneous modality gap through coarse, global feature-level alignment and fusion. 2) The erosion of bit independence and consequent limitations on the semantic representation capacity of hash codes during feature-level hash code learning. 3) The insufficiency of binary label-based pairwise semantic preservation strategies in capturing intricate fine-grained semantic correlations within multi-modal data. To address these challenges, this paper introduces the Dynamic Bit-wise Semantic Transformer Hashing (DBSTH) framework. Remarkably, it treats each hash bit as a unique semantic concept, facilitating concept-level alignment of heterogeneous modalities. This safeguards bit independence and augments representation capabilities. Specifically, we devise a dynamic unit fusion strategy for the adaptive combination of local multi-modal information units, facilitating the acquisition of bit-wise semantic concepts. Subsequently, we incorporate a transformer encoder to refine these concepts by uncovering latent correlations among distinct concepts. Finally, we perform the multi-modal alignment and fusion on the fine-grained concept-level, independently encoding each concept to its corresponding hash bit. To provide enhanced guidance for concept learning, a label prototype learning mechanism is introduced, which learns prototype embeddings for all categories through the consideration of co-occurrence priors. This mechanism effectively captures fine-grained explicit semantic correlations and generates supervising hash codes. Additionally, to improve the robustness of the hashing model in handling noisy multi-modal data, a masked concept learning strategy is introduced, facilitating the acquisition of resilient semantic concepts. Extensive experiments conducted on three widely tested multi-modal retrieval datasets demonstrate the superiority of our method in conventional, noisy, and open-set retrieval scenarios.
- 7 November 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 48 (3) , 2954-2969
Dynamic Bit-Wise Semantic Transformer Hashing for Multi-Modal Retrieval
Multi-modal hashing aims to succinctly encode heterogeneous modalities into binary hash codes, facilitating efficient multimedia retrieval characterized by low storage demands and high retrieval speed. Despite the commendable achievements of existing methods, they still face three crucial challenges: 1) Inadequate bridging of the heterogeneous modality gap through coarse, global feature-level alignment and fusion. 2) The erosion of bit independence and consequent limitations on the semantic representation capacity of hash codes during feature-level hash code learning. 3) The insufficiency of binary label-based pairwise semantic preservation strategies in capturing intricate fine-grained semantic correlations within multi-modal data. To address these challenges, this paper introduces the Dynamic Bit-wise Semantic Transformer Hashing (DBSTH) framework. Remarkably, it treats each hash bit as a unique semantic concept, facilitating concept-level alignment of heterogeneous modalities. This safeguards bit independence and augments representation capabilities. Specifically, we devise a dynamic unit fusion strategy for the adaptive combination of local multi-modal information units, facilitating the acquisition of bit-wise semantic concepts. Subsequently, we incorporate a transformer encoder to refine these concepts by uncovering latent correlations among distinct concepts. Finally, we perform the multi-modal alignment and fusion on the fine-grained concept-level, independently encoding each concept to its corresponding hash bit. To provide enhanced guidance for concept learning, a label prototype learning mechanism is introduced, which learns prototype embeddings for all categories through the consideration of co-occurrence priors. This mechanism effectively captures fine-grained explicit semantic correlations and generates supervising hash codes. Additionally, to improve the robustness of the hashing model in handling noisy multi-modal data, a masked concept learning strategy is introduced, facilitating the acquisition of resilient semantic concepts. Extensive experiments conducted on three widely tested multi-modal retrieval datasets demonstrate the superiority of our method in conventional, noisy, and open-set retrieval scenarios.
- 3 October 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 38, 814-818
Application of Mathematical Models in Forecasting the Usage of Water, Electricity and Gas in Hospital Logistics
To explore the prediction effect of mathematical models on water, electricity and gas usage in hospital logistics operations, in order to obtain the best prediction model and improve the level of hospital logistics refinement management. Monthly data of water, electricity and gas usage of hospital logistics from January 2016 to December 2020 were selected for the training and testing of the model, a nonlinear self-organizing regression neural network model and an exponential smoothing correlation model were constructed, and the prediction effect was evaluated and analyzed by evaluation indexes. The goodness-of-fit of the NARNN model reached 0.87047 on the training data set of water usage, 0.91166 on the validation set, and 0.94479 on the test set, with an overall model goodness-of-fit value of 0.88075, and an overall goodness-of-fit value of 0.95808, close to 1, in the prediction of gas usage, and the seasonal multiplication model in the water and electricity The RMSEs of the seasonal multiplication model in both water and electricity usage forecasting reached the lowest. All the mathematical models used in this paper can effectively predict the water, electricity and gas usage data. By evaluating the index results, a comprehensive comparison shows that the nonlinear self-organizing neural network model has the best prediction effect and can provide prediction data support for information decision making and energy saving and emission reduction deployment in the hospital logistics department.
- 3 October 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 38, 814-818
Application of Mathematical Models in Forecasting the Usage of Water, Electricity and Gas in Hospital Logistics
To explore the prediction effect of mathematical models on water, electricity and gas usage in hospital logistics operations, in order to obtain the best prediction model and improve the level of hospital logistics refinement management. Monthly data of water, electricity and gas usage of hospital logistics from January 2016 to December 2020 were selected for the training and testing of the model, a nonlinear self-organizing regression neural network model and an exponential smoothing correlation model were constructed, and the prediction effect was evaluated and analyzed by evaluation indexes. The goodness-of-fit of the NARNN model reached 0.87047 on the training data set of water usage, 0.91166 on the validation set, and 0.94479 on the test set, with an overall model goodness-of-fit value of 0.88075, and an overall goodness-of-fit value of 0.95808, close to 1, in the prediction of gas usage, and the seasonal multiplication model in the water and electricity The RMSEs of the seasonal multiplication model in both water and electricity usage forecasting reached the lowest. All the mathematical models used in this paper can effectively predict the water, electricity and gas usage data. By evaluating the index results, a comprehensive comparison shows that the nonlinear self-organizing neural network model has the best prediction effect and can provide prediction data support for information decision making and energy saving and emission reduction deployment in the hospital logistics department.