ArCHer: Training Language Model Agents via Hierarchical Multi-Turn RL

Yifei Zhou, Andrea Zanette, Jiayi Pan, Sergey Levine, Aviral Kumar
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:62178-62209, 2024.

Abstract

Large language models (LLMs) have the potential to tackle sequential decision-making problems due to their generalist capabilities. Instead of optimizing “myopic” surrogate objectives such as human preferences within a single turn, in such problems, we wish to directly optimize long-term objectives, such as user satisfaction over an entire dialogue with an LLM or delayed success metrics in web navigation. Multi-turn reinforcement learning (RL) provides an appealing approach to directly optimize long-term objectives, but how can we design effective and efficient multi-turn RL algorithms for LLMs? In this work, we propose an algorithmic framework to multi-turn RL for LLMs that preserves the flexibility of token-by-token RL used in single-turn RL problems, while still accommodating long horizons and delayed rewards more effectively. Our framework, the Actor-Critic Framework with a Hierarchical Structure (ArCHer), combines a high-level off-policy RL algorithm that trains a value function with a low-level RL algorithm that trains a token-by-token policy. While ArCHer can be instantiated with multiple RL algorithms, a particularly convenient instantiation is to use temporal difference (TD) learning at the high level and on-policy token-level policy gradient at the low level. Empirically, we show that ArCHer significantly improves efficiency and performance of multi-turn LLM tasks, attaining sample efficiency boosts of about 100x over prior on-policy methods and converging to a much better performance than other off-policy methods.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-zhou24t, title = {{A}r{CH}er: Training Language Model Agents via Hierarchical Multi-Turn {RL}}, author = {Zhou, Yifei and Zanette, Andrea and Pan, Jiayi and Levine, Sergey and Kumar, Aviral}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {62178--62209}, year = {2024}, editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix}, volume = {235}, series = {Proceedings of Machine Learning Research}, month = {21--27 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v235/main/assets/zhou24t/zhou24t.pdf}, url = {https://proceedings.mlr.press/v235/zhou24t.html}, abstract = {Large language models (LLMs) have the potential to tackle sequential decision-making problems due to their generalist capabilities. Instead of optimizing “myopic” surrogate objectives such as human preferences within a single turn, in such problems, we wish to directly optimize long-term objectives, such as user satisfaction over an entire dialogue with an LLM or delayed success metrics in web navigation. Multi-turn reinforcement learning (RL) provides an appealing approach to directly optimize long-term objectives, but how can we design effective and efficient multi-turn RL algorithms for LLMs? In this work, we propose an algorithmic framework to multi-turn RL for LLMs that preserves the flexibility of token-by-token RL used in single-turn RL problems, while still accommodating long horizons and delayed rewards more effectively. Our framework, the Actor-Critic Framework with a Hierarchical Structure (ArCHer), combines a high-level off-policy RL algorithm that trains a value function with a low-level RL algorithm that trains a token-by-token policy. While ArCHer can be instantiated with multiple RL algorithms, a particularly convenient instantiation is to use temporal difference (TD) learning at the high level and on-policy token-level policy gradient at the low level. Empirically, we show that ArCHer significantly improves efficiency and performance of multi-turn LLM tasks, attaining sample efficiency boosts of about 100x over prior on-policy methods and converging to a much better performance than other off-policy methods.} }
Endnote
%0 Conference Paper %T ArCHer: Training Language Model Agents via Hierarchical Multi-Turn RL %A Yifei Zhou %A Andrea Zanette %A Jiayi Pan %A Sergey Levine %A Aviral Kumar %B Proceedings of the 41st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Ruslan Salakhutdinov %E Zico Kolter %E Katherine Heller %E Adrian Weller %E Nuria Oliver %E Jonathan Scarlett %E Felix Berkenkamp %F pmlr-v235-zhou24t %I PMLR %P 62178--62209 %U https://proceedings.mlr.press/v235/zhou24t.html %V 235 %X Large language models (LLMs) have the potential to tackle sequential decision-making problems due to their generalist capabilities. Instead of optimizing “myopic” surrogate objectives such as human preferences within a single turn, in such problems, we wish to directly optimize long-term objectives, such as user satisfaction over an entire dialogue with an LLM or delayed success metrics in web navigation. Multi-turn reinforcement learning (RL) provides an appealing approach to directly optimize long-term objectives, but how can we design effective and efficient multi-turn RL algorithms for LLMs? In this work, we propose an algorithmic framework to multi-turn RL for LLMs that preserves the flexibility of token-by-token RL used in single-turn RL problems, while still accommodating long horizons and delayed rewards more effectively. Our framework, the Actor-Critic Framework with a Hierarchical Structure (ArCHer), combines a high-level off-policy RL algorithm that trains a value function with a low-level RL algorithm that trains a token-by-token policy. While ArCHer can be instantiated with multiple RL algorithms, a particularly convenient instantiation is to use temporal difference (TD) learning at the high level and on-policy token-level policy gradient at the low level. Empirically, we show that ArCHer significantly improves efficiency and performance of multi-turn LLM tasks, attaining sample efficiency boosts of about 100x over prior on-policy methods and converging to a much better performance than other off-policy methods.
APA
Zhou, Y., Zanette, A., Pan, J., Levine, S. & Kumar, A.. (2024). ArCHer: Training Language Model Agents via Hierarchical Multi-Turn RL. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:62178-62209 Available from https://proceedings.mlr.press/v235/zhou24t.html.

Related Material