Surgical Fine-Tuning Improves Adaptation to Distribution Shifts

ICLR 2023(2022)

引用 82|浏览223
暂无评分
摘要
A common approach to transfer learning under distribution shift is to fine-tune the last few layers of a pre-trained model, preserving learned features while also adapting to the new task. This paper shows that in such settings, selectively fine-tuning a subset of layers (which we term surgical fine-tuning) matches or outperforms commonly used fine-tuning approaches. Moreover, the type of distribution shift influences which subset is more effective to tune: for example, for image corruptions, fine-tuning only the first few layers works best. We validate our findings systematically across seven real-world data tasks spanning three types of distribution shifts. Theoretically, we prove that for two-layer neural networks in an idealized setting, first-layer tuning can outperform fine-tuning all layers. Intuitively, fine-tuning more parameters on a small target dataset can cause information learned during pre-training to be forgotten, and the relevant information depends on the type of shift.
更多
查看译文
关键词
Transfer learning,fine-tuning,parameter freezing,distortion of pre-trained models
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络