Fine-Tuning Pre-Trained Language Models Effectively by Optimizing Subnetworks Adaptively (2211.01642v1)

Abstract: Large-scale pre-trained LLMs have achieved impressive results on a wide range of downstream tasks recently. However, fine-tuning an extremely large-scale pre-trained LLM on limited target datasets is often plagued by overfitting and representation degradation. In this paper, we propose a Dynamic Parameter Selection (DPS) algorithm for the large-scale pre-trained models during fine-tuning, which adaptively selects a more promising subnetwork to perform staging updates based on gradients of back-propagation. Experiments on the GLUE benchmark show that DPS outperforms previous fine-tuning methods in terms of overall performance and stability, and consistently achieves better results with variable pre-trained LLMs. In addition, DPS brings a large magnitude of improvement in out-of-domain transferring experiments and low-resource scenarios, which shows that it can maintain stable general contextual features and reduce the representation collapse. We release our code at https://github.com/ZhangHaojie077/DPS