Towards Full-Text Scholarly Argument Mining using Entity-Anchored Claim Extraction
This work introduces a novel approach that leverages scientific entity mentions to identify and structure claims in scholarly literature to address the annotation bottleneck and employs a majority-voting ensemble of large language models to generate a silver-standard dataset comprising 1,487 claims.
Bhuvanesh Verma, Andy Lüking, Mounika Marreddy et al.
· 0 citations