Sample Efficient Reinforcement Learning via Large Vision Language Model Distillation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909612776620032 |
|---|---|
| author | Lee, Donghoon Luu, Tung M. Lee, Younghwan Yoo, Chang D. |
| author_facet | Lee, Donghoon Luu, Tung M. Lee, Younghwan Yoo, Chang D. |
| contents | Recent research highlights the potential of multimodal foundation models in tackling complex decision-making challenges. However, their large parameters make real-world deployment resource-intensive and often impractical for constrained systems. Reinforcement learning (RL) shows promise for task-specific agents but suffers from high sample complexity, limiting practical applications. To address these challenges, we introduce LVLM to Policy (LVLM2P), a novel framework that distills knowledge from large vision-language models (LVLM) into more efficient RL agents. Our approach leverages the LVLM as a teacher, providing instructional actions based on trajectories collected by the RL agent, which helps reduce less meaningful exploration in the early stages of learning, thereby significantly accelerating the agent's learning progress. Additionally, by leveraging the LVLM to suggest actions directly from visual observations, we eliminate the need for manual textual descriptors of the environment, enhancing applicability across diverse tasks. Experiments show that LVLM2P significantly enhances the sample efficiency of baseline RL algorithms. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_11221 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Sample Efficient Reinforcement Learning via Large Vision Language Model Distillation Lee, Donghoon Luu, Tung M. Lee, Younghwan Yoo, Chang D. Machine Learning Recent research highlights the potential of multimodal foundation models in tackling complex decision-making challenges. However, their large parameters make real-world deployment resource-intensive and often impractical for constrained systems. Reinforcement learning (RL) shows promise for task-specific agents but suffers from high sample complexity, limiting practical applications. To address these challenges, we introduce LVLM to Policy (LVLM2P), a novel framework that distills knowledge from large vision-language models (LVLM) into more efficient RL agents. Our approach leverages the LVLM as a teacher, providing instructional actions based on trajectories collected by the RL agent, which helps reduce less meaningful exploration in the early stages of learning, thereby significantly accelerating the agent's learning progress. Additionally, by leveraging the LVLM to suggest actions directly from visual observations, we eliminate the need for manual textual descriptors of the environment, enhancing applicability across diverse tasks. Experiments show that LVLM2P significantly enhances the sample efficiency of baseline RL algorithms. |
| title | Sample Efficient Reinforcement Learning via Large Vision Language Model Distillation |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2505.11221 |