NoisyTune: A Little Noise Can Help You Finetune Pretrained Language Models Better
This addresses the issue of suboptimal fine-tuning for language models in NLP tasks, though it is incremental as it builds on existing fine-tuning methods.
The paper tackles the problem of pretrained language models overfitting to pretraining data and underperforming on downstream tasks by proposing NoisyTune, a method that adds noise to model parameters before fine-tuning, resulting in consistent performance improvements across GLUE and XTREME benchmarks.
Effectively finetuning pretrained language models (PLMs) is critical for their success in downstream tasks. However, PLMs may have risks in overfitting the pretraining tasks and data, which usually have gap with the target downstream tasks. Such gap may be difficult for existing PLM finetuning methods to overcome and lead to suboptimal performance. In this paper, we propose a very simple yet effective method named NoisyTune to help better finetune PLMs on downstream tasks by adding some noise to the parameters of PLMs before fine-tuning. More specifically, we propose a matrix-wise perturbing method which adds different uniform noises to different parameter matrices based on their standard deviations. In this way, the varied characteristics of different types of parameters in PLMs can be considered. Extensive experiments on both GLUE English benchmark and XTREME multilingual benchmark show NoisyTune can consistently empower the finetuning of different PLMs on different downstream tasks.