Language Models as Zero-Shot Planners: Extracting Actionable Knowledge for Embodied Agents

Wenlong Huang, Pieter Abbeel, Deepak Pathak, Igor Mordatch
Proceedings of the 39th International Conference on Machine Learning, PMLR 162:9118-9147, 2022.

Abstract

Can world knowledge learned by large language models (LLMs) be used to act in interactive environments? In this paper, we investigate the possibility of grounding high-level tasks, expressed in natural language (e.g. “make breakfast”), to a chosen set of actionable steps (e.g. “open fridge”). While prior work focused on learning from explicit step-by-step examples of how to act, we surprisingly find that if pre-trained LMs are large enough and prompted appropriately, they can effectively decompose high-level tasks into mid-level plans without any further training. However, the plans produced naively by LLMs often cannot map precisely to admissible actions. We propose a procedure that conditions on existing demonstrations and semantically translates the plans to admissible actions. Our evaluation in the recent VirtualHome environment shows that the resulting method substantially improves executability over the LLM baseline. The conducted human evaluation reveals a trade-off between executability and correctness but shows a promising sign towards extracting actionable knowledge from language models.

Cite this Paper

BibTeX
@InProceedings{pmlr-v162-huang22a, title = {Language Models as Zero-Shot Planners: Extracting Actionable Knowledge for Embodied Agents}, author = {Huang, Wenlong and Abbeel, Pieter and Pathak, Deepak and Mordatch, Igor}, booktitle = {Proceedings of the 39th International Conference on Machine Learning}, pages = {9118--9147}, year = {2022}, editor = {Chaudhuri, Kamalika and Jegelka, Stefanie and Song, Le and Szepesvari, Csaba and Niu, Gang and Sabato, Sivan}, volume = {162}, series = {Proceedings of Machine Learning Research}, month = {17--23 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v162/huang22a/huang22a.pdf}, url = {https://proceedings.mlr.press/v162/huang22a.html}, abstract = {Can world knowledge learned by large language models (LLMs) be used to act in interactive environments? In this paper, we investigate the possibility of grounding high-level tasks, expressed in natural language (e.g. “make breakfast”), to a chosen set of actionable steps (e.g. “open fridge”). While prior work focused on learning from explicit step-by-step examples of how to act, we surprisingly find that if pre-trained LMs are large enough and prompted appropriately, they can effectively decompose high-level tasks into mid-level plans without any further training. However, the plans produced naively by LLMs often cannot map precisely to admissible actions. We propose a procedure that conditions on existing demonstrations and semantically translates the plans to admissible actions. Our evaluation in the recent VirtualHome environment shows that the resulting method substantially improves executability over the LLM baseline. The conducted human evaluation reveals a trade-off between executability and correctness but shows a promising sign towards extracting actionable knowledge from language models.} }
Endnote
%0 Conference Paper %T Language Models as Zero-Shot Planners: Extracting Actionable Knowledge for Embodied Agents %A Wenlong Huang %A Pieter Abbeel %A Deepak Pathak %A Igor Mordatch %B Proceedings of the 39th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2022 %E Kamalika Chaudhuri %E Stefanie Jegelka %E Le Song %E Csaba Szepesvari %E Gang Niu %E Sivan Sabato %F pmlr-v162-huang22a %I PMLR %P 9118--9147 %U https://proceedings.mlr.press/v162/huang22a.html %V 162 %X Can world knowledge learned by large language models (LLMs) be used to act in interactive environments? In this paper, we investigate the possibility of grounding high-level tasks, expressed in natural language (e.g. “make breakfast”), to a chosen set of actionable steps (e.g. “open fridge”). While prior work focused on learning from explicit step-by-step examples of how to act, we surprisingly find that if pre-trained LMs are large enough and prompted appropriately, they can effectively decompose high-level tasks into mid-level plans without any further training. However, the plans produced naively by LLMs often cannot map precisely to admissible actions. We propose a procedure that conditions on existing demonstrations and semantically translates the plans to admissible actions. Our evaluation in the recent VirtualHome environment shows that the resulting method substantially improves executability over the LLM baseline. The conducted human evaluation reveals a trade-off between executability and correctness but shows a promising sign towards extracting actionable knowledge from language models.
APA
Huang, W., Abbeel, P., Pathak, D. & Mordatch, I.. (2022). Language Models as Zero-Shot Planners: Extracting Actionable Knowledge for Embodied Agents. Proceedings of the 39th International Conference on Machine Learning, in Proceedings of Machine Learning Research 162:9118-9147 Available from https://proceedings.mlr.press/v162/huang22a.html.

Related Material