Adaptive Random Feature Regularization on Fine-tuning Deep Neural Networks

TL;DR

This paper introduces AdaRand, a simple and efficient regularization method for fine-tuning deep neural networks that adaptively aligns feature vectors with random references, improving performance without extra source data or heavy computation.

Contribution

AdaRand is a novel regularization technique that adaptively adjusts feature distributions during fine-tuning without auxiliary source information or high computational costs.

Findings

AdaRand outperforms existing regularization methods in fine-tuning tasks.

It effectively reduces overfitting on small target datasets.

The method dynamically updates feature distributions for better class separation.

Abstract

While fine-tuning is a de facto standard method for training deep neural networks, it still suffers from overfitting when using small target datasets. Previous methods improve fine-tuning performance by maintaining knowledge of the source datasets or introducing regularization terms such as contrastive loss. However, these methods require auxiliary source information (e.g., source labels or datasets) or heavy additional computations. In this paper, we propose a simple method called adaptive random feature regularization (AdaRand). AdaRand helps the feature extractors of training models to adaptively change the distribution of feature vectors for downstream classification tasks without auxiliary source information and with reasonable computation costs. To this end, AdaRand minimizes the gap between feature vectors and random reference vectors that are sampled from class conditional Gaussian distributions. Furthermore, AdaRand dynamically updates the conditional distribution to follow the currently updated feature extractors and balance the distance between classes in feature spaces. Our experiments show that AdaRand outperforms the other fine-tuning regularization, which requires auxiliary source information and heavy computation costs.