The reusability of state-of-the-art Pre-trained Language Models (PLMs) is often limited by their generalization problem, where their performance drastically decreases when evaluated on examples that differ from the training dataset, known as Out-of-Distribution (OOD)/unseen examples. This limitation arises from PLMs' reliance on spurious correlations, which work well for frequent example types but not for general examples. To address this issue, we propose a training approach called Mask-tuning, which integrates Masked Language Modeling (MLM) training objectives into the fine-tuning process to enhance PLMs' generalization. Comprehensive experiments demonstrate that Mask-tuning surpasses current state-of-the-art techniques and enhances PLMs' generalization on OOD datasets while improving their performance on in-distribution datasets. The findings suggest that Mask-tuning improves the reusability of PLMs on unseen data, making them more practical and effective for real-world applications.
@article{arxiv.2307.10457,
title = {Improving the Reusability of Pre-trained Language Models in Real-world Applications},
author = {Somayeh Ghanbarzadeh and Hamid Palangi and Yan Huang and Radames Cruz Moreno and Hamed Khanpour},
journal= {arXiv preprint arXiv:2307.10457},
year = {2023}
}
Comments
Accepted as a long paper and awarded as the BEST Resaerch Paper in IEEE IRI'23 (IEEE 24th International conference on Information Reuse and Integrationfor Data Science)