To Learn or Not to Learn: Replaced Token Detection for Learning the Meaning of Negation

Gunjan Bhattarai, Katrin Erk

Abstract
State-of-the-art language models perform well on a variety of language tasks, but they continue to struggle with understanding negation cues in tasks like natural language inference (NLI). Inspired by Hossain et al. (2020), who show under-representation of negation in language model pretraining datasets, we experiment with additional pretraining with negation data for which we introduce two new datasets. We also introduce a new learning strategy for negation building on ELECTRA’s (Clark et al., 2020) replaced token detection objective. We find that continuing to pretrain ELECTRA-Small’s discriminator leads to substantial gains on a variant of RTE (Recognizing Textual Entailment) with additional negation. On SNLI (Stanford NLI) (Bowman et al., 2015), there are no gains due to the extreme under-representation of negation in the data. Finally, on MNLI (Multi-NLI) (Williams et al., 2018), we find that performance on negation cues is primarily stymied by neutral-labeled examples.
Anthology ID:
2024.lrec-main.1411
Volume:
Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)
Month:
May
Year:
2024
Address:
Torino, Italia
Editors:
Nicoletta Calzolari, Min-Yen Kan, Veronique Hoste, Alessandro Lenci, Sakriani Sakti, Nianwen Xue
Venues:
LREC | COLING
SIG:
Publisher:
ELRA and ICCL
Note:
Pages:
16237–16250
Language:
URL:
https://aclanthology.org/2024.lrec-main.1411
DOI:
Bibkey:
Cite (ACL):
Gunjan Bhattarai and Katrin Erk. 2024. To Learn or Not to Learn: Replaced Token Detection for Learning the Meaning of Negation. In Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024), pages 16237–16250, Torino, Italia. ELRA and ICCL.
Cite (Informal):
To Learn or Not to Learn: Replaced Token Detection for Learning the Meaning of Negation (Bhattarai & Erk, LREC-COLING 2024)
Copy Citation:
PDF:
https://preview.aclanthology.org/nschneid-patch-4/2024.lrec-main.1411.pdf
PDF Search