Surgical Fine-Tuning Improves Adaptation to Distribution Shifts
read the original abstract
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.
This paper has not been read by Pith yet.
Forward citations
Cited by 5 Pith papers
-
Sample-wise Targeted Adversarial Attacks on Test-time Adaptation
Proposes meta-learning attack with priority-aware gradient alignment for sample-wise targeted attacks on TTA that maintain label distribution consistency with no-attack baseline.
-
Selective, Regularized, and Calibrated: Harnessing Vision Foundation Models for Cross-Domain Few-Shot Semantic Segmentation
HERA is a select-regularize-calibrate framework adapting frozen vision foundation models for cross-domain few-shot semantic segmentation via hierarchical layer selection with ETR, prior-guided regularization, and pixe...
-
Intermediate Representations are Strong AI-Generated Image Detectors
Intermediate layer embedding sensitivity to perturbations distinguishes AI-generated images from real ones, yielding higher AUROC on GenImage and Forensics Small benchmarks than prior methods.
-
Task-agnostic Low-rank Residual Adaptation for Efficient Federated Continual Fine-Tuning
Fed-TaLoRA uses task-agnostic low-rank residual adaptation with post-aggregation calibration to enable efficient federated continual fine-tuning across sequential tasks under non-IID conditions.
-
Generalizable Deepfake Detection Based on Forgery-aware Layer Masking and Multi-artifact Subspace Decomposition
FMSD improves cross-dataset generalization in deepfake detection by using gradient-based layer masking to select forgery-sensitive weights and SVD to split them into preserved semantic and multiple learnable artifact ...
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.