思维状态机：利用过往推理轨迹增强问题求解能力
State Machine of Thoughts: Leveraging Past Reasoning Trajectories for Enhancing Problem Solving
AI总结：
针对大语言模型智能体推理轨迹复用效率低的问题，提出SMoT框架，用状态机记录过往推理的成败经验，选择最优子解，在24点游戏和出租车导航强化学习任务中显著提升问题求解能力。
摘要：
当前基于大语言模型的智能体在探索-评估框架内进行推理，以树状方式推进问题求解过程。但这类方法往往在问题解决后就忽略了成功的推理轨迹，导致这些轨迹无法被高效用于后续同类问题。为解决这一低效问题，我们采用状态机记录从过往推理轨迹中提取的经验。在状态机中，状态代表分解后的子问题，状态转移则反映子问题间的依赖关系，同时记录成功与失败的轨迹。借助状态机中的经验，我们提出的State Machine of Thoughts（SMoT，思维状态机）能够选择最优子解并规避错误子解。实验表明，SMoT可在24点游戏和出租车导航强化学习游戏这两类高探索需求问题中显著提升问题求解能力。
英文摘要：
Current Large Language Model-based agents reason within an exploration-evaluation framework, navigating problem-solving processes in a tree-like manner. However, these methods often neglect successful reasoning trajectories once a problem is resolved, leading to inefficient use of these trajectories for future analogous problems. To address this inefficiency, we adopt a state machine to record experience derived from previous reasoning trajectories. Within the state machine, states represent decomposed sub-problems, while state transitions reflect the dependencies among sub-problems. The state machine records both successful and failed trajectories. Utilizing the experience from the state machine, our proposed State Machine of Thoughts (SMoT) selects the most optimal sub-solutions and avoids incorrect ones. Our experiments show that SMoT can significantly improve problem-solving abilities in two exploration-intensive problems: the 24-point game and a taxi navigation reinforcement learning game.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
