Adding UD-based hypergraph similarity to the attention heads of multilingual PLMs improves paraphrase identification by about 3.9% accuracy and 6.1% F1 on average.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Evaluating the Effectiveness of Linguistic Knowledge in Pretrained Language Models: A Case Study of Universal Dependencies
Adding UD-based hypergraph similarity to the attention heads of multilingual PLMs improves paraphrase identification by about 3.9% accuracy and 6.1% F1 on average.