超越单一负偏好：面向以LLM为中心的历史实体链接的多负DPO

中文摘要 AI 辅助

大型语言模型（LLMs）近来在历史实体链接任务中展现出潜力，但该任务的偏好优化通常仅针对每个训练实例使用一个负候选。这丢弃了为同一提及所检索到的其余候选信息。我们提出了多负直接偏好优化（MDPO），这是一种基于参考的成对目标，它将正确实体与每个提及关联的所有有效拒绝候选进行比较。MDPO保留了DPO的Bradley-Terry公式，同时通过掩蔽的、长度归一化的序列分数利用完整候选集。我们在hipe-2020和newseye上评估MDPO，涵盖法语、德语、英语、瑞典语和芬兰语的历史报纸文本。实验表明，MDPO优于监督微调和单负DPO，尤其在NIL提及、语义歧义、OCR噪声和历史困难名称方面有显著提升。进一步分析区分了候选生成和选择错误，表明候选检索仍是端到端实体链接的关键瓶颈。这些结果证明，纳入实例内所有负候选是基于LLM的历史实体链接的一种简单而有效的改进。

英文摘要

Large language models (LLMs) have recently shown promise for historical entity linking, but preference optimization for this task is often formulated with only one negative candidate per training instance. This discards information from the remaining candidates retrieved for the same mention. We introduce multi-negative direct preference optimisation (MDPO), a reference-based pairwise objective that compares the correct entity with all valid rejected candidates associated with each mention. MDPO preserves the Bradley-Terry formulation of DPO while exploiting the complete candidate set through masked, length-normalised sequence scores. We evaluate MDPO on hipe-2020 and newseye, covering French, German, English, Swedish, and Finnish historical newspaper text. Experiments show that MDPO improves over supervised fine-tuning and single-negative DPO, with particularly strong gains for NIL mentions, semantic ambiguity, OCR noise, and historically difficult names. Further analyses disentangle candidate-generation and selection errors, showing that candidate retrieval remains a key bottleneck for end-to-end entity linking. These results demonstrate that incorporating all within-instance negative candidates is a simple and effective improvement for LLM-based historical entity linking.

发表机构

• University of La Rochelle（拉罗谢尔大学）
• University of Toulouse（图卢兹大学）
• University of Innsbruck（因斯布鲁克大学）
• University of Ljubljana（卢布尔雅那大学）

机构由 AI 辅助整理，请以论文原文为准。