arXiv cs.CL· MiHyeon Kim, Ji-Eun Han, Eunkyeong Lee, Jeehyun Lim, Hwijung Ryu·· 11 小时前AI 评分20
SciGraphSelect：通过知识-推理图实现可审计的科学数据选择
SciGraphSelect: Auditable Scientific Data Selection through Knowledge--Reasoning Graphs
AI 导读
SciGraphSelect 是一个基于参考依据、可审计的科学数据选择框架，将领域知识、可复用推理操作与参考佐证关系组织为类型化知识-推理图，再用概念覆盖、关系与路径共支持及方向对齐对候选文档打分。该框架在参考侧保留丰富关系结构，对候选侧证据保守处理、不假设关系蕴含，从而为每次选择暴露可追溯的证据。论文被 NeurIPS 2026 AI4Science Workshop 接收。
正文 · 原文
Abstract:Scientific data selection is typically driven by relevance or utility scores, providing limited insight into why a document is retained. We introduce SciGraphSelect, a reference-grounded framework for auditable scientific data selection. SciGraphSelect organizes domain-specific knowledge, reusable reasoning operations, and reference-attested relations into a typed knowledge--reasoning graph, then scores candidate documents using concept coverage, relation and path co-support, and directional alignment. To avoid over-interpreting document embeddings, rich relational structure is preserved on the reference side while candidate-side evidence is treated conservatively without assuming relation entailment. By exposing the reference-grounded evidence behind each selection, SciGraphSelect brings traceability to the upstream construction of scientific training corpora and complements downstream scientific verification.
| Comments: | 40th Conference on Neural Information Processing Systems (NeurIPS 2026). Workshop: AI4Science |
| Subjects: | Artificial Intelligence (cs.AI); Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.41280 [cs.AI] |
| (or arXiv:2610.41280v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.41280 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Mihyeon Kim [view email]
[v1]
Fri, 30 Oct 2026 05:36:49 UTC (1,146 KB)
来源：arXiv cs.CL · arxiv.org