English

Task-guided Disentangled Tuning for Pretrained Language Models

Computation and Language 2022-03-23 v1

Abstract

Pretrained language models (PLMs) trained on large-scale unlabeled corpus are typically fine-tuned on task-specific downstream datasets, which have produced state-of-the-art results on various NLP tasks. However, the data discrepancy issue in domain and scale makes fine-tuning fail to efficiently capture task-specific patterns, especially in the low data regime. To address this issue, we propose Task-guided Disentangled Tuning (TDT) for PLMs, which enhances the generalization of representations by disentangling task-relevant signals from the entangled representations. For a given task, we introduce a learnable confidence model to detect indicative guidance from context, and further propose a disentangled regularization to mitigate the over-reliance problem. Experimental results on GLUE and CLUE benchmarks show that TDT gives consistently better results than fine-tuning with different PLMs, and extensive analysis demonstrates the effectiveness and robustness of our method. Code is available at https://github.com/lemon0830/TDT.

Keywords

Cite

@article{arxiv.2203.11431,
  title  = {Task-guided Disentangled Tuning for Pretrained Language Models},
  author = {Jiali Zeng and Yufan Jiang and Shuangzhi Wu and Yongjing Yin and Mu Li},
  journal= {arXiv preprint arXiv:2203.11431},
  year   = {2022}
}

Comments

Findings of ACL 2022

R2 v1 2026-06-24T10:21:24.896Z