A*-Thought：在低资源设置下通过双向压缩实现高效推理
A*-Thought: Efficient Reasoning via Bidirectional Compression for Low-Resource Settings
AI总结：
本文提出 A*-Thought 框架，用 A* 树搜索、路径代价函数和双向重要性估计压缩高价值推理链，在低预算下显著提升 QwQ-32B 性能并在高预算下减少近半 token。
摘要：
大型推理模型（Large Reasoning Models，LRM）通过延长思维长度取得了卓越性能。然而，冗长的思维轨迹会导致效率下降。现有大多数方法受制于过度思考这一假设，试图通过压缩思维链来进行高效推理，但这往往会造成性能下降。为解决该问题，我们提出 A*-Thought，这是一个基于树搜索的高效统一框架，旨在从这些模型生成的大量推理链中识别并分离出最关键的思维。它将 LRM 的推理过程表述为一棵搜索树，其中每个节点表示巨大推理空间中的一个推理片段。通过将 A* 搜索算法与针对推理路径的代价函数相结合，该框架能够高效压缩思维链，并确定一条具有高信息密度和低成本的推理路径。此外，我们还提出了一种双向重要性估计机制，进一步细化这一搜索过程，并使其效率超越均匀采样。在若干高级数学任务上的大量实验表明，A*-Thought 在巨大的搜索空间中有效平衡了性能与效率。具体而言，在低预算条件下，A*-Thought 可将 QwQ-32B 的性能提升 2.39 倍；在高预算条件下，可将输出 token 长度减少近 50%。所提出的方法还兼容其他若干 LRM，证明了其泛化能力。代码可在 https://github.com/AI9Stars/AStar-Thought 获取。
英文摘要：
Large Reasoning Models (LRMs) achieve superior performance by extending the thought length. However, a lengthy thinking trajectory leads to reduced efficiency. Most of the existing methods are stuck in the assumption of overthinking and attempt to reason efficiently by compressing the Chain-of-Thought, but this often leads to performance degradation. To address this problem, we introduce A*-Thought, an efficient tree search-based unified framework designed to identify and isolate the most essential thoughts from the extensive reasoning chains produced by these models. It formulates the reasoning process of LRMs as a search tree, where each node represents a reasoning span in the giant reasoning space. By combining the A* search algorithm with a cost function specific to the reasoning path, it can efficiently compress the chain of thought and determine a reasoning path with high information density and low cost. In addition, we also propose a bidirectional importance estimation mechanism, which further refines this search process and enhances its efficiency beyond uniform sampling. Extensive experiments on several advanced math tasks show that A*-Thought effectively balances performance and efficiency over a huge search space. Specifically, A*-Thought can improve the performance of QwQ-32B by 2.39$\times$ with low-budget and reduce the length of the output token by nearly 50% with high-budget. The proposed method is also compatible with several other LRMs, demonstrating its generalization capability. The code can be accessed at: https://github.com/AI9Stars/AStar-Thought.
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。
