01 / MOTIVATION
Why this work matters
Scientific evidence can be difficult to retrieve from the informal wording of social-media claims. The work investigates how candidate retrieval and language-model reranking can prioritize better scientific sources.
A scientific-source retrieval system that combines initial retrieval with LLM-based reranking for social-media claims.
02 / METHODOLOGY
Research approach
Candidate scientific documents are retrieved broadly, then an LLM-based reranker selects sources that more closely support checking the claim.
03 / CONTRIBUTIONS
Main contribution
Scientific evidence retrieval strengthened by LLM reranking.
Consult the complete publication for numerical findings and experimental qualifications.
04 / PUBLICATION
Read the paper
Mohotarema Rashid, Nansu Baniya, Anirban Saha Anik, Xiaoying Song and Lingzi Hong. (2026). SciClaimSeekers at CheckThat! 2026: Retrieving Scientific Sources for Social Media Claims with LLM Reranking. Working Notes of CLEF 2026, CEUR Workshop Proceedings, Vol. 4283, pp. 1184–1191.