Fair Evaluation in Concept Normalization: a Large-scale Comparative Analysis for BERT-based Models

3777

Knowledge Triples