arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2601.04210cs.CLcs.AIcs.ITmath.IT
收藏

无复杂度偏见的思维递归分解

Complexity Agnostic Recursive Decomposition of Thoughts

Southwest Jiaotong University(西南交通大学);Southwestern University of Finance and Economics(西南财经大学)

Kaleem Ullah Qasim, Jiashu Zhang, Hafiz Saif Ur Rehman

AI总结:

CARD通过预估问题复杂度并递归分解提升多步推理的准确性和效率

摘要:

大型语言模型在多步推理中常常失败,因为其固定推理策略忽略了问题特定的难度。我们引入CARD(无复杂度偏见的递归分解)框架,在生成前预测问题复杂度并相应调整分解。我们的系统包括MRCE(多维推理复杂度估计器),一个0.6B Qwen模型,能够从问题文本中预测30个细粒度特征,并一个两阶段递归求解器:(1)基于任务概况进行层次分解为K步;(2)通过递归MRCE分析分配每一步的思维预算(1、5-9或10个思维)。在三个推理模型(Qwen3-0.6B、DeepSeek-R1-Distill-Qwen-1.5B、Qwen3-1.7B)上进行评估,CARD在GSM8K上达到81.4%至89.2%的准确率,比固定分解基线减少1.88倍至2.40倍的token成本。在MATH-500上,CARD使用1.71倍至5.74倍更少的token达到75.1至86.8%的准确率。我们的结果表明,预先估计复杂度能够实现更高的准确性和显著的效率提升。

英文摘要:

Large language models often fail on multi-step reasoning due to fixed reasoning strategies that ignore problem specific difficulty. We introduce CARD (Complexity Agnostic Recursive Decomposition), a framework that predicts problem complexity before generation and adapts decomposition accordingly. Our system comprises MRCE (Multi-dimensional Reasoning Complexity Estimator), a 0.6B Qwen model predicting 30 fine-grained features from question text and a two-stage recursive solver: (1) hierarchical decomposition into K steps based on task profile and (2) per-step thought budget allocation (1, 5-9, or 10 thoughts) via recursive MRCE profiling. Evaluated on three reasoning models (Qwen3-0.6B, DeepSeek-R1-Distill-Qwen-1.5B, Qwen3-1.7B), CARD achieves 81.4% to 89.2% accuracy on GSM8K while reducing token cost by 1.88x to 2.40x compared to fixed decomposition baselines. On MATH-500, CARD reaches 75.1 to 86.8% accuracy using 1.71x to 5.74x fewer tokens. Our results demonstrate that preemptive complexity estimation enables both higher accuracy and significant efficiency gains.

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。

↑