NTCIR-19 SciClaimEval 任务概览：科学论文中的跨模态论断验证

View PDF HTML (experimental)

Abstract:This paper presents an overview of \emph{SciClaimEval}, a pilot task at NTCIR-19 on verifying scientific claims against the tables and figures of the papers in which they appear. Our dataset pairs authentic claims from 180 papers in machine learning, natural language processing, and medicine with contrastive versions of their tables and figures, one supporting the claim and one minimally modified so that it does not. The task comprises two subtasks: deciding whether a piece of evidence supports a claim, and selecting which of two evidence versions does. Twenty-two teams registered, the highest number among NTCIR-19 tasks, and 14 submitted 177 runs. Most systems built on multimodal large language models, relying on prompting and multi-stage inference rather than task-specific fine-tuning. On the first subtask, the best systems reached 98.1\% pair accuracy, but all six top-ranked systems relied on the contrastive pairs; the best system that judged each claim-evidence pair on its own scored 24.3 points lower. Open-weight models, embedded in carefully designed pipelines, matched proprietary ones. Yet verifying a single claim against its evidence, the setting that matters in practice, remains an open problem.

Comments:	The first two authors contributed equally. More information about the shared task is available at this https URL
Subjects:	Computation and Language (cs.CL)
Cite as:	arXiv:2611.06090 [cs.CL]
	(or arXiv:2611.06090v1 [cs.CL] for this version)
	https://doi.org/10.48550/arXiv.2611.06090 arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Xanh Ho Thi [view email]
[v1] Wed, 4 Nov 2026 07:33:52 UTC (920 KB)