AdaFilter: Adaptive Filter Fine-tuning for Deep Transfer Learning

TL;DR

AdaFilter is an adaptive fine-tuning method that selectively updates convolutional filters in pre-trained neural networks on a per-example basis, improving transfer learning performance.

Contribution

This paper introduces AdaFilter, a novel adaptive fine-tuning approach that dynamically selects filters for each example, addressing limitations of standard fine-tuning methods.

Findings

Reduces average classification error by 2.54% across datasets

Outperforms standard fine-tuning methods in experiments

Demonstrates effectiveness on 7 public image datasets

Abstract

There is an increasing number of pre-trained deep neural network models. However, it is still unclear how to effectively use these models for a new task. Transfer learning, which aims to transfer knowledge from source tasks to a target task, is an effective solution to this problem. Fine-tuning is a popular transfer learning technique for deep neural networks where a few rounds of training are applied to the parameters of a pre-trained model to adapt them to a new task. Despite its popularity, in this paper, we show that fine-tuning suffers from several drawbacks. We propose an adaptive fine-tuning approach, called AdaFilter, which selects only a part of the convolutional filters in the pre-trained model to optimize on a per-example basis. We use a recurrent gated network to selectively fine-tune convolutional filters based on the activations of the previous layer. We experiment with 7 public image classification datasets and the results show that AdaFilter can reduce the average classification error of the standard fine-tuning by 2.54%.