Adaptive Parametric Prototype Learning for Cross-Domain Few-Shot Classification

TL;DR

This paper introduces a novel adaptive parametric prototype learning method for cross-domain few-shot classification, addressing domain shifts with meta-learning, prototype regularization, and transductive fine-tuning, achieving superior results on benchmarks.

Contribution

We propose a new adaptive parametric prototype learning approach that learns class prototypes from concatenated features and employs meta-learning and transductive fine-tuning for better cross-domain few-shot classification.

Findings

APPL outperforms state-of-the-art methods on multiple benchmarks.

Prototype regularization improves class discrimination.

Transductive fine-tuning enhances adaptation to target domain.

Abstract

Cross-domain few-shot classification induces a much more challenging problem than its in-domain counterpart due to the existence of domain shifts between the training and test tasks. In this paper, we develop a novel Adaptive Parametric Prototype Learning (APPL) method under the meta-learning convention for cross-domain few-shot classification. Different from existing prototypical few-shot methods that use the averages of support instances to calculate the class prototypes, we propose to learn class prototypes from the concatenated features of the support set in a parametric fashion and meta-learn the model by enforcing prototype-based regularization on the query set. In addition, we fine-tune the model in the target domain in a transductive manner using a weighted-moving-average self-training approach on the query instances. We conduct experiments on multiple cross-domain few-shot benchmark datasets. The empirical results demonstrate that APPL yields superior performance than many state-of-the-art cross-domain few-shot learning methods.