Binary Token-Level Classification with DeBERTa for All-Type MWE Identification: A Lightweight Approach with Linguistic Enhancement
Fuente:
arXiv
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917225125904384 |
|---|---|
| author | Rossini, Diego van der Plas, Lonneke |
| author_facet | Rossini, Diego van der Plas, Lonneke |
| contents | We present a comprehensive approach for multiword expression (MWE) identification that combines binary token-level classification, linguistic feature integration, and data augmentation. Our DeBERTa-v3-large model achieves 69.8% F1 on the CoAM dataset, surpassing the best results (Qwen-72B, 57.8% F1) on this dataset by 12 points while using 165x fewer parameters. We achieve this performance by (1) reformulating detection as binary token-level START/END/INSIDE classification rather than span-based prediction, (2) incorporating NP chunking and dependency features that help discontinuous and NOUN-type MWEs identification, and (3) applying oversampling that addresses severe class imbalance in the training data. We confirm the generalization of our method on the STREUSLE dataset, achieving 78.9% F1. These results demonstrate that carefully designed smaller models can substantially outperform LLMs on structured NLP tasks, with important implications for resource-constrained deployments. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_19360 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Binary Token-Level Classification with DeBERTa for All-Type MWE Identification: A Lightweight Approach with Linguistic Enhancement Rossini, Diego van der Plas, Lonneke Computation and Language We present a comprehensive approach for multiword expression (MWE) identification that combines binary token-level classification, linguistic feature integration, and data augmentation. Our DeBERTa-v3-large model achieves 69.8% F1 on the CoAM dataset, surpassing the best results (Qwen-72B, 57.8% F1) on this dataset by 12 points while using 165x fewer parameters. We achieve this performance by (1) reformulating detection as binary token-level START/END/INSIDE classification rather than span-based prediction, (2) incorporating NP chunking and dependency features that help discontinuous and NOUN-type MWEs identification, and (3) applying oversampling that addresses severe class imbalance in the training data. We confirm the generalization of our method on the STREUSLE dataset, achieving 78.9% F1. These results demonstrate that carefully designed smaller models can substantially outperform LLMs on structured NLP tasks, with important implications for resource-constrained deployments. |
| title | Binary Token-Level Classification with DeBERTa for All-Type MWE Identification: A Lightweight Approach with Linguistic Enhancement |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2601.19360 |