Task-Prior Conditional Variational Auto-Encoder for Few-Shot Image Classification

TL;DR

This paper introduces TP-VAE, a novel conditional variational auto-encoder that effectively handles nonuniform few-shot image classification by leveraging task-level prior regularization, outperforming existing methods.

Contribution

The paper proposes TP-VAE, a new model that addresses nonuniform class sample sizes in few-shot learning, improving performance over state-of-the-art methods.

Findings

High performance in nonuniform few-shot scenarios

Outperforms state-of-the-art in standard scenarios

1-shot accuracy increased by about 3%

Abstract

Transductive methods always outperform inductive methods in few-shot image classification scenarios. However, the existing few-shot methods contain a latent condition: the number of samples in each class is the same, which may be unrealistic. To cope with those cases where the query shots of each class are nonuniform (i.e. nonuniform few-shot learning), we propose a Task-Prior Conditional Variational Auto-Encoder model named TP-VAE, conditioned on support shots and constrained by a task-level prior regularization. Our method obtains high performance in the more challenging nonuniform few-shot scenarios. Moreover, our method outperforms the state-of-the-art in a wide range of standard few-shot image classification scenarios. Among them, the accuracy of 1-shot increased by about 3\%.