COGRE 与 HIT@DICT：自生成奖励实现可解释关系抽取

View PDF HTML (experimental)

Abstract:Despite the remarkable reasoning capabilities of large language models, they still struggle with one-shot relation extraction without predefined relation labels. We identify two pitfalls: models are often misled by irrelevant tokens instead of relation-conveying semantics, and they often fail to align with the abstraction level human annotators expect. We introduce a novel framework that closes this gap with two components: (1) COGRE, a cognitively-inspired reasoning framework that structures RE into a series of processes mimicking human text-processing; and (2) HIT@DICT, a reinforcement learning intermediate reward strategy that encourages reasoning to align with relational labels by rewarding relation-relevant phrases in reasoning. The reward is derived on a credit dictionary automatically extracted from correct predictions. Our experiments show that our framework improves both accuracy and explanation quality by addressing these two pitfalls. For example, COGRE with Qwen2.5-14B-Instruct on One-shot NYT29 achieves 24.65% F1, surpassing prior reasoning-based designs. Optimizing this approach with RL using HIT@DICT further improves performance by +23.46% points. Finally, human evaluation shows that our best model generates relational phrases closely aligned with gold labels, increasing human explanation quality ratings by 54% (relative).

Subjects:	Computation and Language (cs.CL); Information Retrieval (cs.IR)
Cite as:	arXiv:2510.06198 [cs.CL]
	(or arXiv:2510.06198v4 [cs.CL] for this version)
	https://doi.org/10.48550/arXiv.2510.06198 arXiv-issued DOI via DataCite

Submission history

From: Zhengliang Shi [view email]
[v1] Tue, 7 Oct 2025 17:53:55 UTC (1,034 KB)
[v2] Mon, 16 Feb 2026 14:42:00 UTC (1,013 KB)
[v3] Sat, 13 Jun 2026 04:01:29 UTC (995 KB)
[v4] Wed, 21 Oct 2026 08:05:58 UTC (989 KB)