Scalable Offline Model-Based RL with Action Chunks
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_ | 1866912754996084736 |
|---|---|
| author | Park, Kwanyoung Park, Seohong Lee, Youngwoon Levine, Sergey |
| author_facet | Park, Kwanyoung Park, Seohong Lee, Youngwoon Levine, Sergey |
| contents | In this paper, we study whether model-based reinforcement learning (RL), in particular model-based value expansion, can provide a scalable recipe for tackling complex, long-horizon tasks in offline RL. Model-based value expansion fits an on-policy value function using length-n imaginary rollouts generated by the current policy and a learned dynamics model. While larger n reduces bias in value bootstrapping, it amplifies accumulated model errors over long horizons, degrading future predictions. We address this trade-off with an \emph{action-chunk} model that predicts a future state from a sequence of actions (an "action chunk") instead of a single action, which reduces compounding errors. In addition, instead of directly training a policy to maximize rewards, we employ rejection sampling from an expressive behavioral action-chunk policy, which prevents model exploitation from out-of-distribution actions. We call this recipe \textbf{Model-Based RL with Action Chunks (MAC)}. Through experiments on highly challenging tasks with large-scale datasets of up to 100M transitions, we show that MAC achieves the best performance among offline model-based RL algorithms, especially on challenging long-horizon tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_08108 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Scalable Offline Model-Based RL with Action Chunks Park, Kwanyoung Park, Seohong Lee, Youngwoon Levine, Sergey Machine Learning Artificial Intelligence In this paper, we study whether model-based reinforcement learning (RL), in particular model-based value expansion, can provide a scalable recipe for tackling complex, long-horizon tasks in offline RL. Model-based value expansion fits an on-policy value function using length-n imaginary rollouts generated by the current policy and a learned dynamics model. While larger n reduces bias in value bootstrapping, it amplifies accumulated model errors over long horizons, degrading future predictions. We address this trade-off with an \emph{action-chunk} model that predicts a future state from a sequence of actions (an "action chunk") instead of a single action, which reduces compounding errors. In addition, instead of directly training a policy to maximize rewards, we employ rejection sampling from an expressive behavioral action-chunk policy, which prevents model exploitation from out-of-distribution actions. We call this recipe \textbf{Model-Based RL with Action Chunks (MAC)}. Through experiments on highly challenging tasks with large-scale datasets of up to 100M transitions, we show that MAC achieves the best performance among offline model-based RL algorithms, especially on challenging long-horizon tasks. |
| title | Scalable Offline Model-Based RL with Action Chunks |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2512.08108 |