AI

Generative Ontology Induction: 大規模言語モデルを用いた文書コーパスからのドメイン非依存型スキーマ発見

Generative Ontology Induction: Domain-Agnostic Schema Discovery from Document Corpora Using Large Language Models

arxiv2026年7月21日

日本語要約

本論文は、LLMを用いて文書コーパスからドメイン非依存型のスキーマを発見するGenerative Ontology Inductionを提案する。これにより、知識構造化が自動化され、AIが非構造化情報を理解・整理する能力が向上する可能性がある。

English Summary

This paper introduces Generative Ontology Induction, a method using LLMs to discover domain-agnostic schemas from document corpora. This could automate knowledge structuring and enhance AI's ability to understand and organize unstructured information.


元記事を読む𝕏 でシェア
Generative Ontology Induction: 大規模言語モデルを用いた文書コーパスからのドメイン非依存型スキーマ発見 | Mirai Signal