PRACT: Optimizing Principled Reasoning and Acting of LLM Agent

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Liu, Zhiwei, Yao, Weiran, Zhang, Jianguo, Murthy, Rithesh, Yang, Liangwei, Liu, Zuxin, Lan, Tian, Zhu, Ming, Tan, Juntao, Kokane, Shirley, Hoang, Thai, Niebles, Juan Carlos, Heinecke, Shelby, Wang, Huan, Savarese, Silvio, Xiong, Caiming
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929556798046208
author Liu, Zhiwei
Yao, Weiran
Zhang, Jianguo
Murthy, Rithesh
Yang, Liangwei
Liu, Zuxin
Lan, Tian
Zhu, Ming
Tan, Juntao
Kokane, Shirley
Hoang, Thai
Niebles, Juan Carlos
Heinecke, Shelby
Wang, Huan
Savarese, Silvio
Xiong, Caiming
author_facet Liu, Zhiwei
Yao, Weiran
Zhang, Jianguo
Murthy, Rithesh
Yang, Liangwei
Liu, Zuxin
Lan, Tian
Zhu, Ming
Tan, Juntao
Kokane, Shirley
Hoang, Thai
Niebles, Juan Carlos
Heinecke, Shelby
Wang, Huan
Savarese, Silvio
Xiong, Caiming
contents We introduce the Principled Reasoning and Acting (PRAct) framework, a novel method for learning and enforcing action principles from trajectory data. Central to our approach is the use of text gradients from a reflection and optimization engine to derive these action principles. To adapt action principles to specific task requirements, we propose a new optimization framework, Reflective Principle Optimization (RPO). After execution, RPO employs a reflector to critique current action principles and an optimizer to update them accordingly. We develop the RPO framework under two scenarios: Reward-RPO, which uses environmental rewards for reflection, and Self-RPO, which conducts self-reflection without external rewards. Additionally, two RPO methods, RPO-Traj and RPO-Batch, is introduced to adapt to different settings. Experimental results across four environments demonstrate that the PRAct agent, leveraging the RPO framework, effectively learns and applies action principles to enhance performance.
format Preprint
id arxiv_https___arxiv_org_abs_2410_18528
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle PRACT: Optimizing Principled Reasoning and Acting of LLM Agent
Liu, Zhiwei
Yao, Weiran
Zhang, Jianguo
Murthy, Rithesh
Yang, Liangwei
Liu, Zuxin
Lan, Tian
Zhu, Ming
Tan, Juntao
Kokane, Shirley
Hoang, Thai
Niebles, Juan Carlos
Heinecke, Shelby
Wang, Huan
Savarese, Silvio
Xiong, Caiming
Artificial Intelligence
We introduce the Principled Reasoning and Acting (PRAct) framework, a novel method for learning and enforcing action principles from trajectory data. Central to our approach is the use of text gradients from a reflection and optimization engine to derive these action principles. To adapt action principles to specific task requirements, we propose a new optimization framework, Reflective Principle Optimization (RPO). After execution, RPO employs a reflector to critique current action principles and an optimizer to update them accordingly. We develop the RPO framework under two scenarios: Reward-RPO, which uses environmental rewards for reflection, and Self-RPO, which conducts self-reflection without external rewards. Additionally, two RPO methods, RPO-Traj and RPO-Batch, is introduced to adapt to different settings. Experimental results across four environments demonstrate that the PRAct agent, leveraging the RPO framework, effectively learns and applies action principles to enhance performance.
title PRACT: Optimizing Principled Reasoning and Acting of LLM Agent
topic Artificial Intelligence
url https://arxiv.org/abs/2410.18528