教师简介

---

国家级青年人才、博导

王可泽博士分别于2012年和2017年在中山大学获得学士和博士学位，随后前往美国加州大学洛杉矶分校做博士后研究员。他于2019年获得香港理工大学哲学博士学位。他从博士阶段开始探究如何减少深度学习对训练样本的依赖和如何从海量无标签数据挖掘出有价值信息，提出“引导-自步-协同”长效自主学习等基础学习范式，引入伪标签学习机制，以融合领域知识和语义信息的深度表达学习为主线，从感知到认知，从监督学习到自监督学习，再到自主学习，通过认知启发和引导的因果推理技术，逐步构建面向多模态大模型的视觉计算与推理理论及方法体系。相关工作在国际顶级学术期刊和会议发表论文50余篇，含一篇Cell子刊，以及IEEE PAMI、TNNLS、IJCV、TIP、TMM、TCSVT，以及CVPR、ICCV等领域顶级会议论文30余篇，Google学术被引用累计3300余次，单篇最高引用924次，有四篇论文被评为ESI高被引论文。他获得吴文俊人工智能自然科学奖二等奖，人工智能学会CAAI优秀博士论文奖，国际知名学术评估机构AI 2000最有影响力学者提名奖。

 

研究领域

---

因果认知驱动的复杂视觉推理、多智能体协同高阶推理、多模态生产式AI、具身智能

 

研究目标

---

近年来，以大模型为核心的人工智能技术蓬勃发展，在语言理解、图像识别、多模态交互等诸多领域取得了令人瞩目的成就。然而，当前的大模型在通用性、可解释性、可靠性等关键方面依然存在明显不足，这不仅制约了其在更广泛场景中的落地应用，也对人工智能的长远发展提出了前所未有的挑战。面对这一时代挑战，我们将以理论创新和应用实践双轮驱动，致力于创造全新的多模态大模型学习范式与推理策略，并在现有基础上实现颠覆性改进。我们的目标是大幅提升大模型的通用性与鲁棒性，使其不仅能够跨模态高效理解与表达，还能具备因果发现、逻辑推理与举一反三等类人高级认知能力，从而为人工智能迈向自主学习、自我进化的新阶段奠定坚实基础。我们坚信这不仅是一项科研任务，更是一场面向未来的技术使命——我们希望推动人工智能从“强大工具”走向“智慧伙伴”，更好地为人类造福。

 

海外经历

---

美国加州大学洛杉矶分校博士后研究员

 

获奖及荣誉

---

2022 AI 200最有影响力学者提名奖

2020 Volunteer highlight of IEEE Transactions on Pattern Analysis and Machine Intelligence 

2019 中国人工智能学会优秀博士学位论文奖(每年最多评10名)

2018 吴文俊人工智能自然科学奖（排名第二）

2016 国家博士研究生奖学金(排名前1%)

2015 国家博士研究生奖学金(排名前1%)

 

主要学术兼职

---

担任国际顶级会议Association for Computational Linguistics (ACL) 的领域主席（Area Chair）

担任国际知名期刊Image and Vision Computing的执行编辑

担任国际知名期刊The Visual Computer的副编辑

担任国际知名期刊The Journal of Visual Communication and Image Representation的副编辑

担任以下期刊的审稿人:

– IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI)

– IEEE Transactions on Neural Networks and Learning Systems (TNNLS)

– Applied Soft Computing Journal

– IEEE Transactions on Circuits and Systems for Video Technology (TCSVT)

– IEEE Transactions on Image Processing (TIP)

– IEEE Transactions on Multimedia (TMM)

– Pattern Recognition (PR)

– Neural Networks

– Neurocomputing

 • 担任以下会议的审稿人:

– International Conference on Computer Vision and Pattern Recognition (CVPR)

– European Conference on Computer Vision (ECCV) 

– Neural Information Processing Systems (NeurIPS) 

– AAAI Conference on Artificial Intelligence (AAAI) 

– IEEE International Conference on Computer Vision (ICCV) 

– International Joint Conference on Artificial Intelligence (IJCAI) 

– IEEE International Conference on Robotics and Automation (ICRA) 

– Winter Conference on Applications of Computer Vision (WACV) 

– Asian Conference on Computer Vision (ACCV) 

– International Conference on Pattern Recognition (ICPR) 

 

代表性论著

---

 注：(*)表示通讯作者，(+)表示共同作者

[1] Ziyi Tang, Zechuan Chen, Jiarui Yang, Jiayao Mai, Yongsen Zheng, Keze Wang, Jinrui Chen, Liang Lin. AlphaAgent: LLM-Driven Alpha Mining with Regularized Exploration to Counteract Alpha Decay. In KDD 2025.

[2] Jusheng Zhang, Zimeng Huang, Yijia Fan, Ningyuan Liu, Mingyan Li, Zhuojie Yang, Jiawei Yao, Jian Wang, Keze Wang*. KABB: Knowledge-Aware Bayesian Bandits for Dynamic Expert Coordination in Multi-Agent Systems. In ICML 2025.

[3] Zeqing Wang, Qingyang Ma, Wentao Wan, Haojie Li, Keze Wang*. Yonghong Tian. Is this Generated Person Existed in Real-world? Fine-grained Detecting and Calibrating Abnormal Human-body. In CVPR 2025.

[4]  Wentao Wan, Zhuojie Yang, Yongcan Chen, Chenglin Luo, Ruilin Wang, Kehao Cai, Nan Kang, Liang Lin, Keze Wang*. SR-FoT: A Syllogistic-Reasoning Framework of Thought for Large Language Models Tackling Knowledge-based Reasoning Tasks. In AAAI 2025.

[5] Hefeng Wu, Hao Jiang, Keze Wang*, Ziyi Tang, Xianghuan He, Liang Lin, Improving Network Interpretability via Explanation Consistency Evaluation. In IEEE Transactions on Multimedia, 2024.

[6] Qingyi Liu, Jinhui Qin, Wenxuan Ye, Hao Mou, Yuxuan He, Keze Wang*. Adaptive Prompt Routing for Arbitrary Text Style Transfer with Pre-trained Language Models. In AAAI 2024.

[7] Linsheng Chen, Guangrun Wang, Liuchun Yuan, Keze Wang*, Ken Deng, Philip H.S. Torr. NeRF-VPT: Learning Novel View Representations with Neural Radiance Fields via View Prompt Tuning. In AAAI 2024.

[8] Hui Fu, Zeqing Wang, Ke Gong, Keze Wang, Tianshui Chen, Haojie Li, Haifeng Zeng, Wenxiong Kang. Mimic: Speaking Style Disentanglement for Speech-Driven 3D Facial Animation. In AAAI 2024.

[9] Lingling Li, Weicong Li, Qiyuan Ding, Chengpei Tang, Keze Wang*. Gesture Generation via Diffusion Model with Attention Mechanism. In ICASSP 2024.

[10] Junfan Lin, Keze Wang*, Ziliang Chen, Xiaodan Liang, Liang Lin. Towards Causality-Aware Inferring: A Sequential Discriminative Approach for Medical Diagnosis. IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI), 2023.

[11] Xipeng Chen, Junzheng Zhang, Keze Wang, Pengxu Wei, Liang Lin. Multi-Person 3D Pose Estimation With Occlusion Reasoning. In IEEE Transactions on Multimedia, 2022.

[12] Yang Liu, Keze Wang*, Lingbo Liu, Haoyuan Lan, Liang Lin. TCGL: Temporal Contrastive Graph for Self-supervised Video Representation Learning. In IEEE Transactions on Image Processing (T-IP), 2022.

[13] Guangrun Wang, Keze Wang*, Guangcong Wang, Phillip HS Torr, Liang Lin. Solving Inefficiency of Self-supervised Representation Learning. In ICCV 2021.

[14] Arjun Akula+, Keze Wang+, Changsong Liu, Sari Saba-Sadiya, Hongjing Lu, Sinisa Todorovic, Joyce Chai, Song-Chun Zhu*, F-ToM: Explaining with Theory-of-Mind via Fault-Lines for Enhancing Human Trust in Image Recognition Models. In iScience, 2021.

[15] Yang Liu, Keze Wang*, Guanbin Li, Liang Lin. Semantics-aware Adaptive Knowledge Distillation for Sensor-to-Vision Action Recognition. To appear in IEEE Transactions on Image Processing (T-IP), 2021.

[16] Keze Wang, Liang Lin, Chenhan Jiang, Chen Qian, and Pengxu Wei. 3D Human Pose Machines with Self-supervised Learning. In IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI), vol. 42, no. 5, pp. 1069– 1082, 2020.

[17] Junfan Lin, Zhongzhan Huang, Keze Wang*, Xiaodan Liang, Weiwei Chen, Liang Lin. Continuous Transition: Improving Sample Efficiency for Continuous Control Problems via MixUp. In Proc. of International Conference on Robotics and Automation (ICRA), 2021.

[18] Qingxing Cao, Bailin Li, Xiaodan Liang, Keze Wang, Liang Lin*. Knowledge-Routed Visual Question Reasoning: Challenges for Deep Representation Embedding. To appear in IEEE Transactions on Neural Networks and Learning Systems (TNNLS), 2021.

[19] Guangrun Wang, Guangcong Wang, Keze Wang, Xiaodan Liang, Liang Lin, Grammatically Recognizing Images with Tree Convolution. In Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), 2020.

[20] Guangrun Wang, Keze Wang, Liang Lin. Adaptively Connected Neural Networks. In Proc. of IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2019.

[21] Keze Wang, Liang Lin, Xiaopeng Yan, Ziliang Chen, Dongyu Zhang, Lei Zhang. Self-supervised Sample Mining with Switchable Selection Criteria for Object Detection. In IEEE Transactions on Neural Networks and Learning Systems (T-NNLS), vol. 30, no. 3, pp. 834–850, 2019.

[22] Yukai Shi, Guanbin Li, Qingxing Cao, Keze Wang, Liang Lin. Face hallucination by attentive sequence optimization with reinforcement learning. In IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI), vol. 42, no. 11, pp. 2809–2824, 2019.

[23] Hefeng Wu, Yafei Hu, Keze Wang, Hanhui Li, Lin Nie, Hui Cheng. Instance-aware representation learning and association for online multiperson tracking. In Pattern Recognition, vol. 94, pp. 25–34, 2019.

[24] Keze Wang, Liang Lin, Chuangjie Ren, Wei Zhang, Wenxiu Sun. Convolutional Memory Blocks for Depth Data Representation Learning. In Proceedings of the 27th International Joint Conference on Artificial Intelligence and the 23rd European Conference on Artificial Intelligence (IJCAI), 2018.

[25] Keze Wang, Xiaopeng Yan, Dongyu Zhang, Lei Zhang, Liang Lin. Towards Human-Machine Cooperation: Self-supervised Sample Mining for Object Detection. In Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018.

[26] Guanbin Li, Yuan Xie, Tianhao Wei, Keze Wang, Liang Lin. Flow Guided Recurrent Neural Encoder for Video Salient Object Detection. In Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018.

[27] Liang Lin, Keze Wang, Deyu Meng, Wangmeng Zuo, and Lei Zhang. Active Self-Paced Learning for Cost-Effective and Progressive Face Identification. In IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI), vol. 40, no. 1, pp. 7–19, 2018.

[28] Hui Cheng, Zhuoqi Zheng, Jinhao He, Chongyu Chen, Keze Wang, Liang Lin. Embedding Temporally Consistent Depth Recovery for Real-time Dense Mapping in Visual-inertial Odometry. In IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp. 693–698, 2018.

[29] Keze Wang, Dongyu Zhang, Liang Lin, Ya Li, and Ruimao Zhang, Cost-Effective Active Learning for Deep Image Classification. In IEEE Transactions on Circuits and Systems for Video Technology (T-CSVT), vol. 27, no. 12, pp. 2591– 2600, 2017.

[30] Yukai Shi, Keze Wang, Chongyu Chen, Li Xu and Liang Lin. Structure-Preserving Image Superresolution via Contextualized Multi-task Learning. In IEEE Transactions on Multimedia (TMM), vol. 19, no. 12, pp. 2804–2815, 2017.

[31] Ziliang Chen, Keze Wang, Xiao Wang, Pai Peng, and Liang Lin. Deep Co-Space: Sample Mining Across Feature Transformation for Semi-supervised Learning. In IEEE Transactions on Circuits and Systems for Video Technology (TCSVT), 2017.

[32] Mude Lin, Liang Lin, Xiaodan Liang, Keze Wang, and Hui Cheng, Recurrent 3D Pose Sequence Machines. In Proc. of IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017. (oral).

[33] Liang Lin, Keze Wang, Wangmeng Zuo, Meng Wang, Jiebo Luo, Lei Zhang, A Deep Structured Model with Radius–Margin Bound for 3D Human Activity Recognition. In International Journal of Computer Vision (IJCV), 118(2), 256- 273, 2016.

[34] Keze Wang, Liang Lin, Jiangbo Lu, Chenglong Li, Keyang Shi, PISA: Pixelwise Image Saliency by Aggregating Complementary Appearance Contrast Measures with Edge-Preserving Coherence. In IEEE Transactions on Image Processing (TIP), 24(10), 3019-3033, 2015.

[35] Keze Wang, Shengfu Zhai, Hui Cheng, Xiaodan Liang, Liang Lin. Human Pose Estimation from Still Depth Image via Inference Embedded Multi-task Learning. In Proceedings of the ACM International Conference on Multimedia (ACM MM), 2016. (oral, full paper)

[36] Keze Wang, Liang Lin, Wangmeng Zuo, Shuhang Gu, Lei Zhang. Dictionary Pair Classifier Driven Convolutional Neural Networks for Object Detection. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016.

[37] Keze Wang, Xiaolong Wang, Liang Lin, Meng Wang, Wangmeng Zuo, 3D human activity recognition with reconfigurable convolutional neural networks. In Proceedings of the ACM International Conference on Multimedia (ACM MM), pp. 97-106, 2014. (oral, full paper)

[38] Yukai Shi, Keze Wang, Li Xu, Liang Lin, Local and Holistic- Structure Preserving Image Super Resolution via Deep Joint Component Learning. In Proceedings of the IEEE International Conference on Multimedia and Expo (ICME), 2016. (oral)

[39] Linnan Zhu, Keze Wang, Liang Lin, Lei Zhang, Learning a Lightweight Deep Convolutional Network for Joint Age and Gender Recognition. In Proceedings of the IEEE International Conference on Pattern Recognition (ICPR), 2016. (oral)

[40] Keyang Shi, Keze Wang, Jiangbo Lu, Liang Lin, Pisa: Pixelwise image saliency by aggregating complementary appearance contrast measures with spatial priors. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 2115-2122, 2013.