A5049451354- Reinforcement Learning in Robotics
- Artificial Intelligence in Games
- Neural dynamics and brain function
- Advanced Memory and Neural Computing
- Advanced Bandit Algorithms Research
- Blockchain Technology Applications and Security
- Modular Robots and Swarm Intelligence
- Functional Brain Connectivity Studies
- Cryptography and Data Security
- Adversarial Robustness in Machine Learning
- Mobile Crowdsensing and Crowdsourcing
- Receptor Mechanisms and Signaling
- Cloud Data Security Solutions
- Simulation Techniques and Applications
University of Chinese Academy of Sciences
2020-2024
Shandong Institute of Automation
2022-2023
Chinese Academy of Sciences
2020-2023
Beijing Academy of Artificial Intelligence
2022-2023
Jilin Province Science and Technology Department
2018
Jilin University
2018
Opponent modeling is essential to exploit sub-optimal opponents in strategic interactions. Most previous works focus on building explicit models predict the opponents' styles or strategies, which require a large amount of data train model and lack adaptability unknown opponents. In this work, we propose novel Learning Exploit (L2E) framework for implicit opponent modeling. L2E acquires ability through few interactions with different during training neural network can quickly adapt new...
Despite the potential of Multi-Agent Reinforcement Learning (MARL) in addressing numerous complex tasks, training a single team MARL agents to handle multiple diverse tasks remains challenge. In this paper, we introduce novel Multi-task method based on Knowledge Transfer cooperative (MKT-MARL). By learning from task-specific teachers, our approach empowers attain expert-level performance tasks. MKT-MARL utilizes knowledge distillation algorithm specifically designed for multi-agent...
Experience replay plays a crucial role in Reinforcement Learning (RL), enabling the agent to remember and reuse experience from past. Most previous methods sample transitions using simple heuristics like uniformly sampling or prioritizing those good ones. Since humans can learn both bad experiences, more sophisticated algorithms need be developed. Inspired by potential energy physics, this work introduces artificial field into develops Potentialized Replay (PotER) as new effective algorithm...
Multi-task reinforcement learning endeavors to accomplish a set of different tasks with single policy. To enhance data efficiency by sharing parameters across multiple tasks, common practice segments the network into distinct modules and trains routing recombine these task-specific policies. However, existing approaches employ fixed number for all neglecting that varying difficulties commonly require amounts knowledge. This work presents Dynamic Depth Routing (D2R) framework, which learns...
Efficient collaboration in the centralized training with decentralized execution (CTDE) paradigm remains a challenge cooperative multi-agent systems. We identify divergent action tendencies among agents as significant obstacle to CTDE's efficiency, requiring large number of samples achieve unified consensus on agents' policies. This divergence stems from lack adequate team consensus-related guidance signals during credit assignment CTDE. To address this, we propose Intrinsic Action Tendency...
Efficient collaboration in the centralized training with decentralized execution (CTDE) paradigm remains a challenge cooperative multi-agent systems. We identify divergent action tendencies among agents as significant obstacle to CTDE's efficiency, requiring large number of samples achieve unified consensus on agents' policies. This divergence stems from lack adequate team consensus-related guidance signals during credit assignments CTDE. To address this, we propose Intrinsic Action Tendency...
Reinforcement learning (RL) algorithms typically require orders of magnitude more interactions than humans to learn effective policies. Research on memory in neuroscience suggests that humans' efficiency benefits from associating their experiences and reconstructing potential events. Inspired by this finding, we introduce a human brainlike structure for agents build general framework based improve the RL sampling efficiency. Since is similar reconstruction process psychology, name newly...
Multi-task reinforcement learning endeavors to accomplish a set of different tasks with single policy. To enhance data efficiency by sharing parameters across multiple tasks, common practice segments the network into distinct modules and trains routing recombine these task-specific policies. However, existing approaches employ fixed number for all neglecting that varying difficulties commonly require amounts knowledge. This work presents Dynamic Depth Routing (D2R) framework, which learns...
Cloud computing has been developing at a rapid speed, playing an important role in many fields, especially environments like hospitals which produce lot of data every day and have specific users.Because the security information stored cloud cannot be guaranteed, we propose safe storage medical based on attribute encryption.This paper focuses how to apply attribute-based encryption hospitals' environment, design access process different users environment by using encryption.Our goal is build...