Intra-task Mutual Attention based Vision Transformer for Few-Shot Learning

TL;DR

This paper introduces an intra-task mutual attention mechanism in Vision Transformers for few-shot learning, enhancing feature focus and representation sharing between support and query images, leading to improved classification performance.

Contribution

It proposes a novel intra-task mutual attention method with patch token swapping in pre-trained ViT models, combined with self-supervised pre-training and meta-learning fine-tuning, for more effective few-shot classification.

Findings

Achieves superior performance on five few-shot benchmarks.

Reduces the number of parameters needing fine-tuning.

Demonstrates effectiveness and efficiency of the proposed method.

Abstract

Humans possess remarkable ability to accurately classify new, unseen images after being exposed to only a few examples. Such ability stems from their capacity to identify common features shared between new and previously seen images while disregarding distractions such as background variations. However, for artificial neural network models, determining the most relevant features for distinguishing between two images with limited samples presents a challenge. In this paper, we propose an intra-task mutual attention method for few-shot learning, that involves splitting the support and query samples into patches and encoding them using the pre-trained Vision Transformer (ViT) architecture. Specifically, we swap the class (CLS) token and patch tokens between the support and query sets to have the mutual attention, which enables each set to focus on the most useful information. This facilitates the strengthening of intra-class representations and promotes closer proximity between instances of the same class. For implementation, we adopt the ViT-based network architecture and utilize pre-trained model parameters obtained through self-supervision. By leveraging Masked Image Modeling as a self-supervised training task for pre-training, the pre-trained model yields semantically meaningful representations while successfully avoiding supervision collapse. We then employ a meta-learning method to fine-tune the last several layers and CLS token modules. Our strategy significantly reduces the num- ber of parameters that require fine-tuning while effectively uti- lizing the capability of pre-trained model. Extensive experiments show that our framework is simple, effective and computationally efficient, achieving superior performance as compared to the state-of-the-art baselines on five popular few-shot classification benchmarks under the 5-shot and 1-shot scenarios