Detecting Knowledge Inconsistencies Across Text, Tables, and Knowledge Graphs
- 类型:arxiv
- 标识:2607.25959
- 链接:http://arxiv.org/abs/2607.25959v1
- 主分类:rag
- 形态:method
- 被引:0
- 被引来源:Semantic Scholar + OpenAlex
- S2被引:0
- OpenAlex被引:0
- 影响力被引:0
- TLDR:Kontrast is presented, an automatic framework that uses Text-to-SPARQL and LLM reasoning to compare table-based answers with KG evidence and categorize the resulting inconsistencies, and shows that text, tables, and KGs can complement and correct one another through systematic comparison.
- OpenAlex ID:W7171632910
- OpenAlex DOI:10.48550/arxiv.2607.25959
- DOI:10.48550/arxiv.2607.25959
- DOI来源:OpenAlex
- 开放获取:green
- 开放获取链接:https://doi.org/10.48550/arxiv.2607.25959
- OpenAlex更新:2026-08-25
- 待LLM分类:否
- 标题中文:跨文本、表格与知识图谱的知识不一致检测
- TLDR中文:本文提出 Kontrast,一种利用 Text-to-SPARQL 与 LLM 推理将基于表格的答案与 KG 证据进行对比并对所产生的不一致性进行分类的自动框架,并表明文本、表格与 KG 可通过系统性对比相互补充与纠错。
- 来源文件:
- /inbox/tom/_candidates/2026-07-29-agent-rag-longcontext-candidates.json
- /inbox/tom/_candidates/2026-07-30-agent-rag-longcontext-candidates.json
- [S2 enrich]
- [OpenAlex backfill]