边适应边学习：面向科学问题的大语言模型智能工具使用适应
Adapting While Learning: Grounding LLMs for Scientific Problems with Intelligent Tool Usage Adaptation
AI总结：
针对大语言模型在科学问题中的幻觉和工具过度依赖，提出AWL两组件微调方法，通过知识学习和工具使用适应，在六个基准上显著提升准确率并超越先进模型。
摘要：
大语言模型（LLMs）在解决科学问题方面展现出有前景的能力，但常常遭受幻觉问题的困扰。虽然将LLMs与工具集成可以缓解这一问题，但针对工具使用进行微调的模型会过度依赖工具，并产生不必要的成本。受人类专家在选择解决方案前评估问题复杂性的启发，我们提出了一种新颖的两组件微调方法，即“边适应边学习”（AWL）。在第一个组件中，即世界知识学习（WKL），LLMs通过从工具生成的解决方案中学习来内化科学知识。在第二个组件中，即工具使用适应（TUA），我们根据模型的准确率将问题分类为简单或困难，并训练模型对简单问题保持直接推理，而对困难问题则切换到工具。我们在气候科学、流行病学、物理学及其他领域的六个科学基准数据集上验证了我们的方法。与原始指令模型（8B）相比，经过AWL后训练的模型在答案准确率上提高了29.11%，在工具使用准确率上提高了12.72%，甚至在四个自建数据集上超越了包括GPT-4o和Claude-3.5在内的最先进模型。我们的代码已开源，地址为https://github.com/Rose-STL-Lab/Adapting-While-Learning。
英文摘要：
Large Language Models (LLMs) demonstrate promising capabilities in solving scientific problems but often suffer from the issue of hallucination. While integrating LLMs with tools can mitigate this issue, models fine-tuned on tool usage become overreliant on them and incur unnecessary costs. Inspired by how human experts assess problem complexity before selecting solutions, we propose a novel two-component fine-tuning method, Adapting While Learning (AWL). In the first component, World Knowledge Learning (WKL), LLMs internalize scientific knowledge by learning from tool-generated solutions. In the second component, Tool Usage Adaptation (TUA), we categorize problems as easy or hard based on the model's accuracy, and train it to maintain direct reasoning for easy problems while switching to tools for hard ones. We validate our method on six scientific benchmark datasets across climate science, epidemiology, physics, and other domains. Compared to the original instruct model (8B), models post-trained with AWL achieve 29.11% higher answer accuracy and 12.72% better tool usage accuracy, even surpassing state-of-the-art models including GPT-4o and Claude-3.5 on four custom-created datasets. Our code is open-source at https://github.com/Rose-STL-Lab/Adapting-While-Learning.
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。
