Statler: State-Maintaining Language Models for Embodied Reasoning
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2023
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914801220845568 |
|---|---|
| author | Yoneda, Takuma Fang, Jiading Li, Peng Zhang, Huanyu Jiang, Tianchong Lin, Shengjie Picker, Ben Yunis, David Mei, Hongyuan Walter, Matthew R. |
| author_facet | Yoneda, Takuma Fang, Jiading Li, Peng Zhang, Huanyu Jiang, Tianchong Lin, Shengjie Picker, Ben Yunis, David Mei, Hongyuan Walter, Matthew R. |
| contents | There has been a significant research interest in employing large language models to empower intelligent robots with complex reasoning. Existing work focuses on harnessing their abilities to reason about the histories of their actions and observations. In this paper, we explore a new dimension in which large language models may benefit robotics planning. In particular, we propose Statler, a framework in which large language models are prompted to maintain an estimate of the world state, which are often unobservable, and track its transition as new actions are taken. Our framework then conditions each action on the estimate of the current world state. Despite being conceptually simple, our Statler framework significantly outperforms strong competing methods (e.g., Code-as-Policies) on several robot planning tasks. Additionally, it has the potential advantage of scaling up to more challenging long-horizon planning tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2306_17840 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Statler: State-Maintaining Language Models for Embodied Reasoning Yoneda, Takuma Fang, Jiading Li, Peng Zhang, Huanyu Jiang, Tianchong Lin, Shengjie Picker, Ben Yunis, David Mei, Hongyuan Walter, Matthew R. Robotics Computation and Language There has been a significant research interest in employing large language models to empower intelligent robots with complex reasoning. Existing work focuses on harnessing their abilities to reason about the histories of their actions and observations. In this paper, we explore a new dimension in which large language models may benefit robotics planning. In particular, we propose Statler, a framework in which large language models are prompted to maintain an estimate of the world state, which are often unobservable, and track its transition as new actions are taken. Our framework then conditions each action on the estimate of the current world state. Despite being conceptually simple, our Statler framework significantly outperforms strong competing methods (e.g., Code-as-Policies) on several robot planning tasks. Additionally, it has the potential advantage of scaling up to more challenging long-horizon planning tasks. |
| title | Statler: State-Maintaining Language Models for Embodied Reasoning |
| topic | Robotics Computation and Language |
| url | https://arxiv.org/abs/2306.17840 |