citation-verification

To verify academic citations and detect fake references, search the title on Google Scholar and WebSearch, then cross-check the title, authors, and year. You can also query CrossRef, arXiv, or Semantic Scholar APIs to retrieve matching BibTeX metadata and confirm the paper exists.

The best way to audit a bibliography for AI-hallucinated references is to apply a multi-layer validation workflow: verify format, confirm existence via WebSearch, match title/author/year with fuzzy thresholds, and retrieve BibTeX from CrossRef or Semantic Scholar to validate metadata.

You can retrieve BibTeX entries by querying the CrossRef, arXiv, or Semantic Scholar APIs with the paper title or identifier. The verification workflow includes optional Python reference scripts that call these APIs to fetch and validate the BibTeX metadata for your bibliography.

Yes, you can integrate citation verification into ml-paper-writing workflows to enforce proactive checks during academic paper drafting. The process applies WebSearch and Google Scholar existence checks, matching titles, authors, and years with adjustable fuzzy thresholds before adding references.

You need Python dependencies including requests, arxiv, semanticscholar, and bibtexparser to run the optional reference scripts. These libraries enable API calls to fetch metadata and parse BibTeX data for verifying academic references and detecting inaccurate citations.