Accelerating Robotic Reinforcement Learning with Agent Guidance
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912950379347968 |
|---|---|
| author | Chen, Haojun Zou, Zili Ma, Chengdong Pu, Yaoxiang Zhang, Haotong Chen, Yuanpei Yang, Yaodong |
| author_facet | Chen, Haojun Zou, Zili Ma, Chengdong Pu, Yaoxiang Zhang, Haotong Chen, Yuanpei Yang, Yaodong |
| contents | Reinforcement Learning (RL) offers a powerful paradigm for autonomous robots to master generalist manipulation skills through trial-and-error. However, its real-world application is stifled by low sample efficiency. Recent Human-in-the-Loop (HIL) methods accelerate training by using human corrections, yet this approach faces a scalability barrier. Reliance on human supervisors imposes a 1:1 supervision ratio that limits scalability, suffers from operator fatigue over extended sessions, and introduces high variance due to inconsistent human proficiency. We present Agent-guided Policy Search (AGPS), a framework that automates the training pipeline by replacing human supervisors with a multimodal agent. Our key insight is that the agent can be viewed as a semantic world model, injecting intrinsic value priors to structure physical exploration. By using tools, the agent provides precise guidance via corrective waypoints and spatial constraints for exploration pruning. We validate our approach on three tasks, ranging from precision insertion to deformable object manipulation. Results demonstrate that AGPS outperforms HIL methods in sample efficiency. This automates the supervision pipeline, unlocking the path to labor-free and scalable robot learning. Project website: https://agps-rl.github.io/agps/. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_11978 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Accelerating Robotic Reinforcement Learning with Agent Guidance Chen, Haojun Zou, Zili Ma, Chengdong Pu, Yaoxiang Zhang, Haotong Chen, Yuanpei Yang, Yaodong Robotics Artificial Intelligence Reinforcement Learning (RL) offers a powerful paradigm for autonomous robots to master generalist manipulation skills through trial-and-error. However, its real-world application is stifled by low sample efficiency. Recent Human-in-the-Loop (HIL) methods accelerate training by using human corrections, yet this approach faces a scalability barrier. Reliance on human supervisors imposes a 1:1 supervision ratio that limits scalability, suffers from operator fatigue over extended sessions, and introduces high variance due to inconsistent human proficiency. We present Agent-guided Policy Search (AGPS), a framework that automates the training pipeline by replacing human supervisors with a multimodal agent. Our key insight is that the agent can be viewed as a semantic world model, injecting intrinsic value priors to structure physical exploration. By using tools, the agent provides precise guidance via corrective waypoints and spatial constraints for exploration pruning. We validate our approach on three tasks, ranging from precision insertion to deformable object manipulation. Results demonstrate that AGPS outperforms HIL methods in sample efficiency. This automates the supervision pipeline, unlocking the path to labor-free and scalable robot learning. Project website: https://agps-rl.github.io/agps/. |
| title | Accelerating Robotic Reinforcement Learning with Agent Guidance |
| topic | Robotics Artificial Intelligence |
| url | https://arxiv.org/abs/2602.11978 |