arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2406.07080cs.CL
收藏

DARA:用于知识图谱问答的分解-对齐-推理自主语言智能体

DARA: Decomposition-Alignment-Reasoning Autonomous Language Agent for Question Answering over Knowledge Graphs

Technical University of Darmstadt(达姆施塔特工业大学);Hessian Center for AI (hessian.AI)(黑森人工智能中心);Queen’s University(女王大学);Ingenuity Labs Research Institute(Ingenuity 实验室研究院)

Haishuo Fang, Xiaodan Zhu, Iryna Gurevych

AI总结:

本文提出 DARA 框架,通过高层任务分解与低层任务落地将问题解析为形式化查询,并用少量推理轨迹微调 LLM,在零样本 KGQA 基准上超过 GPT-4 上下文学习方案并媲美枚举排序方法。

摘要:

知识图谱问答(KGQA)是自主语言智能体在各类现实应用中良好运行的关键。为提升由大语言模型(LLM)驱动的语言智能体在 KGQA 中的神经符号推理能力,本文提出 Decomposition-Alignment-Reasoning Agent(DARA)框架。DARA 通过双重机制将问题有效解析为形式化查询:高层迭代任务分解与低层任务落地。重要的是,DARA 可利用少量高质量推理轨迹进行高效训练。实验结果表明,在 LLM(如 Llama-2-7B、Mistral)上微调的 DARA,在不同基准的零样本评估中优于基于 GPT-4 的上下文学习智能体以及其他微调智能体,使此类模型更易用于现实应用。本文还表明,DARA 取得了与 KGQA 中最先进的基于枚举与排序的方法相当的性能。

英文摘要:

Answering Questions over Knowledge Graphs (KGQA) is key to well-functioning autonomous language agents in various real-life applications. To improve the neural-symbolic reasoning capabilities of language agents powered by Large Language Models (LLMs) in KGQA, we propose the DecompositionAlignment-Reasoning Agent (DARA) framework. DARA effectively parses questions into formal queries through a dual mechanism: high-level iterative task decomposition and low-level task grounding. Importantly, DARA can be efficiently trained with a small number of high-quality reasoning trajectories. Our experimental results demonstrate that DARA fine-tuned on LLMs (e.g. Llama-2-7B, Mistral) outperforms both in-context learning-based agents with GPT-4 and alternative fine-tuned agents, across different benchmarks in zero-shot evaluation, making such models more accessible for real-life applications. We also show that DARA attains performance comparable to state-of-the-art enumerating-and-ranking-based methods for KGQA.

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。

↑