Policy Adaptation via Language Optimization: Decomposing Tasks for Few-Shot Imitation
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866916555435016192 |
|---|---|
| author | Myers, Vivek Zheng, Bill Chunyuan Mees, Oier Levine, Sergey Fang, Kuan |
| author_facet | Myers, Vivek Zheng, Bill Chunyuan Mees, Oier Levine, Sergey Fang, Kuan |
| contents | Learned language-conditioned robot policies often struggle to effectively adapt to new real-world tasks even when pre-trained across a diverse set of instructions. We propose a novel approach for few-shot adaptation to unseen tasks that exploits the semantic understanding of task decomposition provided by vision-language models (VLMs). Our method, Policy Adaptation via Language Optimization (PALO), combines a handful of demonstrations of a task with proposed language decompositions sampled from a VLM to quickly enable rapid nonparametric adaptation, avoiding the need for a larger fine-tuning dataset. We evaluate PALO on extensive real-world experiments consisting of challenging unseen, long-horizon robot manipulation tasks. We find that PALO is able of consistently complete long-horizon, multi-tier tasks in the real world, outperforming state of the art pre-trained generalist policies, and methods that have access to the same demonstrations. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2408_16228 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Policy Adaptation via Language Optimization: Decomposing Tasks for Few-Shot Imitation Myers, Vivek Zheng, Bill Chunyuan Mees, Oier Levine, Sergey Fang, Kuan Robotics Machine Learning Learned language-conditioned robot policies often struggle to effectively adapt to new real-world tasks even when pre-trained across a diverse set of instructions. We propose a novel approach for few-shot adaptation to unseen tasks that exploits the semantic understanding of task decomposition provided by vision-language models (VLMs). Our method, Policy Adaptation via Language Optimization (PALO), combines a handful of demonstrations of a task with proposed language decompositions sampled from a VLM to quickly enable rapid nonparametric adaptation, avoiding the need for a larger fine-tuning dataset. We evaluate PALO on extensive real-world experiments consisting of challenging unseen, long-horizon robot manipulation tasks. We find that PALO is able of consistently complete long-horizon, multi-tier tasks in the real world, outperforming state of the art pre-trained generalist policies, and methods that have access to the same demonstrations. |
| title | Policy Adaptation via Language Optimization: Decomposing Tasks for Few-Shot Imitation |
| topic | Robotics Machine Learning |
| url | https://arxiv.org/abs/2408.16228 |