About Me
Kai Lv 吕凯
I am an Associate Professor with the School of Computer Science and Technology at Beijing Jiaotong University. My research focuses on learning-based intelligent systems, especially embodied intelligent manipulation, multi-agent collaboration, and re-identification.
My current work studies how agents understand dynamic scenes, coordinate with other agents, and make reliable decisions from visual observations. I am also interested in practical learning systems that connect perception, reasoning, and action.
News
Publication themes- One paper on role-level inductive bias for multi-agent reinforcement learning was accepted by ICML 2026.
- One paper on unsupervised domain adaptation was accepted by IEEE Transactions on Multimedia.
- One paper on reinforcement learning pre-training from videos was accepted by CVPR 2026.
- One paper on active inference for decentralized execution was accepted by AAAI 2026.
- One paper on visual reinforcement learning generalization was accepted by Neural Networks.
- One paper on robust representations for visual reinforcement learning was accepted by ACM Transactions on Multimedia Computing, Communications, and Applications.
- One paper on reinforcement learning pre-training was accepted by ACM Multimedia 2025.
- One paper on continual multi-agent coordination was accepted by IJCAI 2025.
- One paper on offline safe reinforcement learning was accepted by AAMAS 2025.
- Two papers were accepted by AAAI 2025, covering vehicle re-identification and communication delay-tolerant multi-agent collaboration.
- One paper on safe reinforcement learning was accepted by IEEE Transactions on Systems, Man, and Cybernetics: Systems.
Working Experience
Associate Professor
School of Computer Science and Technology, Beijing Jiaotong University
Lecturer
School of Computer Science and Technology, Beijing Jiaotong University
Faculty Postdoctoral Researcher
Beijing Jiaotong University
Selected Publications
中文论文页Embodied Intelligent Manipulation
Visual RLLocal Motion Matters: A Deconstruct-Recompose Paradigm for Reinforcement Learning Pre-training from Videos
The IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2026.
Bidirectional Transition Consistency between Multi-Domain Observations for Visual Reinforcement Learning Generalization
Neural Networks, 2025.
Learning Robust Representations via Bidirectional Transition for Visual Reinforcement Learning
ACM Transactions on Multimedia Computing, Communications and Applications, 2025.
From Pixels to Temporal Correlations: Learning Informative Representations for Reinforcement Learning Pre-training
Proceedings of the 33rd ACM International Conference on Multimedia, 2025.
How to Learn Domain-Invariant Representations for Visual Reinforcement Learning: An Information-Theoretical Perspective
The 33rd International Joint Conference on Artificial Intelligence, 2024.
What Effects the Generalization in Visual Reinforcement Learning: Policy Consistency with Truncated Return Prediction
The 38th AAAI Conference on Artificial Intelligence, 2024.
Building Category Graphs Representation with Spatial and Temporal Attention for Visual Navigation
ACM Transactions on Multimedia Computing, Communications and Applications, 2024.
Agent-Centric Relation Graph for Object Visual Navigation
IEEE Transactions on Circuits and Systems for Video Technology, 2023.
Skill-based Hierarchical Reinforcement Learning for Target Visual Navigation
IEEE Transactions on Multimedia, 2023.
Multi-Agent Collaboration
MARLRole-Level Inductive Bias for Cross-Task Generalization in Multi-Agent Reinforcement Learning
The Forty-third International Conference on Machine Learning, 2026.
Think How Your Teammates Think: Active Inference Can Benefit Decentralized Execution
The 40th AAAI Conference on Artificial Intelligence, 2026.
From General Relation Patterns to Task-Specific Decision-Making in Continual Multi-Agent Coordination
The 34th International Joint Conference on Artificial Intelligence, 2025.
CoDe: Communication Delay-Tolerant Multi-Agent Collaboration via Dual Alignment of Intent and Timeliness
The 39th AAAI Conference on Artificial Intelligence, 2025.
Off-policy Conservative Distributional Reinforcement Learning with Safety Constraints
IEEE Transactions on Systems, Man, and Cybernetics: Systems, 2025.
Enhancing Off-policy Constrained Reinforcement Learning through Adaptive Ensemble C Estimation
The 38th AAAI Conference on Artificial Intelligence, 2024.
Offline Reinforcement Learning with Diffusion-Based Behavior Cloning Term
The 16th International Conference on Knowledge Science, Engineering and Management, 2023.
Lexicographic Actor-Critic Deep Reinforcement Learning for Urban Autonomous Driving
IEEE Transactions on Vehicular Technology, 2022.
A Lightweight and Style-Robust Neural Network for Autonomous Driving in End Side Devices
Connection Science, 2022.
Object Re-Identification
Re-IDLet Confidence Speak: Combining Small and Large Models via Pseudo Labeling for Unsupervised Domain Adaptation
IEEE Transactions on Multimedia, 2026.
Infer the Whole from a Glimpse of a Part: Keypoint-based Knowledge Graph for Vehicle Re-identification
The 39th AAAI Conference on Artificial Intelligence, 2025.
Style Variable and Irrelevant Learning for Generalizable Person Re-identification
ACM Transactions on Multimedia Computing, Communications and Applications, 2024.
Spatially-Regularized Features for Vehicle Re-identification: An Explanation of Where Deep Models Should Focus
IEEE Transactions on Intelligent Transportation Systems, 2023.
Generalization between Different Viewpoints with A Feature Selection Method for Vehicle Re-identification
32nd International Joint Conference on Artificial Intelligence Workshop, 2023.
Pose-Based View Synthesis for Vehicles: A Perspective Aware Method
IEEE Transactions on Image Processing, 2020.
Improving Driver Gaze Prediction with Reinforced Attention
IEEE Transactions on Multimedia, 2020.
Combining Pose Invariant and Discriminative Features for Vehicle Reidentification
IEEE Internet of Things Journal, 2020.
Vehicle Re-Identification with Location and Time Stamps
CVPR Workshops, 2019.