arXivDaily arXiv每日学术速递 周一至周五更新
​通知:联邦节假日调休:哥伦布日 / 原住民日(Columbus Day / Indigenous Peoples' Day),今日(13日)无更新。
arXiv 2405.05008cs.CL

ADELIE:在信息抽取上对齐大型语言模型

ADELIE: Aligning Large Language Models on Information Extraction

Tsinghua University(清华大学)

Yunjia Qi, Hao Peng, Xiaozhi Wang, Bin Xu, Lei Hou, Juanzi Li

AI总结:

针对 LLM 在信息抽取任务中指令遵循不足的问题,本文构建 IEInstruct 对齐语料,并结合指令微调和 DPO 训练 ADELIE,在多类 IE 基准上达到开源 SoTA 且通用能力未明显下降。

摘要:

大型语言模型(LLM)通常在信息抽取(IE)任务上表现不佳,并且难以遵循 IE 任务的复杂指令。这主要源于 LLM 未与人类对齐,因为主流对齐数据集通常不包含 IE 数据。本文提出 ADELIE(Aligning large language moDELs on Information Extraction),这是一种经过对齐的 LLM,可有效解决多种 IE 任务,包括封闭 IE、开放 IE 和按需 IE。我们首先收集并构建用于 IE 的高质量对齐语料库 IEInstruct。随后,我们在 IEInstruct 上使用指令微调训练 ADELIE_SFT。我们进一步使用直接偏好优化(DPO)目标训练 ADELIE_SFT,得到 ADELIE_DPO。在多个留出 IE 数据集上的大量实验表明,我们的模型(ADELIE_SFT 和 ADELIE_DPO)在开源模型中取得了最先进(SoTA)性能。我们进一步探索 ADELIE 的通用能力,实验结果显示其通用能力未出现明显下降。我们将发布代码、数据和模型,以促进进一步研究。

英文摘要:

Large language models (LLMs) usually fall short on information extraction (IE) tasks and struggle to follow the complex instructions of IE tasks. This primarily arises from LLMs not being aligned with humans, as mainstream alignment datasets typically do not include IE data. In this paper, we introduce ADELIE (Aligning large language moDELs on Information Extraction), an aligned LLM that effectively solves various IE tasks, including closed IE, open IE, and on-demand IE. We first collect and construct a high-quality alignment corpus IEInstruct for IE. Then we train ADELIE_SFT using instruction tuning on IEInstruct. We further train ADELIE_SFT with direct preference optimization (DPO) objective, resulting in ADELIE_DPO. Extensive experiments on various held-out IE datasets demonstrate that our models (ADELIE_SFT and ADELIE_DPO) achieve state-of-the-art (SoTA) performance among open-source models. We further explore the general capabilities of ADELIE, and experimental results reveal that their general capabilities do not exhibit a noticeable decline. We will release the code, data, and models to facilitate further research.

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。

↑