提出树状问题框架，提升可分解复杂任务的推理效果
Tree of Problems: Improving structured problem solving with compositionality
- 将复杂问题拆解为相同子任务，构建树状求解结构
- 在多个复杂推理任务上超越CoT、ToT和GoT表现
- 适合需要模块化分步解决的数学与逻辑类问题
大型语言模型（LLMs）通过上下文学习在多种任务中展现出卓越性能。对于需要逐步推理的复杂任务，思维链（CoT）提示已取得显著成果，尤其结合自一致性后效果更佳。然而，部分任务仍难以处理。树思维（ToT）和图思维（GoT）作为替代方案，将复杂问题分解为子问题路径。本文提出树状问题（ToP），一种更简洁的ToT变体，假设其在可分解为相同子任务的复杂任务中表现更优。实验结果表明，该方法在多个复杂推理任务上优于ToT、GoT，且在多数情况下超过CoT。相关代码已公开：https://github.com/ArmelRandy/tree-of-problems。
原文摘要 · Abstract (English)
Large Language Models (LLMs) have demonstrated remarkable performance across multiple tasks through in-context learning. For complex reasoning tasks that require step-by-step thinking, Chain-of-Thought (CoT) prompting has given impressive results, especially when combined with self-consistency. Nonetheless, some tasks remain particularly difficult for LLMs to solve. Tree of Thoughts (ToT) and Graph of Thoughts (GoT) emerged as alternatives, dividing the complex problem into paths of subproblems. In this paper, we propose Tree of Problems (ToP), a simpler version of ToT, which we hypothesise can work better for complex tasks that can be divided into identical subtasks. Our empirical results show that our approach outperforms ToT and GoT, and in addition performs better than CoT on complex reasoning tasks. All code for this paper is publicly available here: https://github.com/ArmelRandy/tree-of-problems.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。