用结构化思维引导大模型解决图优化问题，效果超越更大模型。
GraphThought: Graph Combinatorial Optimization with Thought Generation
- 设计最优思维生成框架，指导大模型产生高质量推理步骤。
- 在GraphArena上超越DeepSeek-V3等更大模型，80亿参数达顶尖性能。
- 适合需要严谨推理的图优化任务，如物流与生物信息学应用。
图组合优化（GCO）问题在物流、生物信息学等领域至关重要。尽管传统求解器占主导地位，大语言模型（LLM）为结构化推理提供了新可能，但在需严格组合分析和多步推演的复杂GCO任务中常出现幻觉推理步骤。本文首次形式化提出最优思维设计（OTD）问题，为生成高质量中间推理步骤提供结构化引导。基于此，我们提出GraphThought框架，通过启发式前向搜索或求解器对齐的后向推理生成有效推理序列。通过在这些结构化思维序列上微调大模型，我们构建了Llama-GT（8B参数），在GraphArena基准上表现优异，显著优于更大的DeepSeek-V3模型。结果表明，仅通过引入结构化推理先验，即可显著提升大模型在GCO任务上的表现，无需增加模型规模。
原文摘要 · Abstract (English)
Graph combinatorial optimization (GCO) problems are central to domains like logistics and bioinformatics. While traditional solvers dominate, large language models (LLMs) offer new possibilities for structured reasoning, yet struggle with complex GCO tasks requiring rigorous combinatorial analysis and multi-step deduction, often producing hallucinated steps. We first formalize the Optimal Thoughts Design (OTD) problem, which provides a structured guidance for producing high-quality intermediate reasoning steps. Building on this formulation, we introduce GraphThought, a novel framework that generates effective reasoning sequences through either heuristic-guided forward search or solver-aligned backward reasoning. By fine-tuning LLMs on these structured thought sequences, we develop Llama-GT, an 8B-parameter model that achieves state-of-the-art performance on the GraphArena benchmark, outperforming significantly larger models like DeepSeek-V3. Our results demonstrate that when scaffolded with structured reasoning priors, principled thought generation can significantly enhance LLM performance on GCO tasks without requiring increased model scale.
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。