Prototypes-oriented Transductive Few-shot Learning with Conditional Transport

TL;DR

This paper introduces PUTM, a novel transductive few-shot learning model that uses conditional transport to address class imbalance by balancing prior distributions of query samples, leading to improved performance.

Contribution

The paper proposes a new imbalanced TFSL model called PUTM that employs conditional transport and an EM-based solver to better handle class imbalance in query samples.

Findings

Outperforms existing methods on four benchmarks

Effectively balances class priors in imbalanced scenarios

Demonstrates superior class-imbalanced generalization

Abstract

Transductive Few-Shot Learning (TFSL) has recently attracted increasing attention since it typically outperforms its inductive peer by leveraging statistics of query samples. However, previous TFSL methods usually encode uniform prior that all the classes within query samples are equally likely, which is biased in imbalanced TFSL and causes severe performance degradation. Given this pivotal issue, in this work, we propose a novel Conditional Transport (CT) based imbalanced TFSL model called {\textbf P}rototypes-oriented {\textbf U}nbiased {\textbf T}ransfer {\textbf M}odel (PUTM) to fully exploit unbiased statistics of imbalanced query samples, which employs forward and backward navigators as transport matrices to balance the prior of query samples per class between uniform and adaptive data-driven distributions. For efficiently transferring statistics learned by CT, we further derive a closed form solution to refine prototypes based on MAP given the learned navigators. The above two steps of discovering and transferring unbiased statistics follow an iterative manner, formulating our EM-based solver. Experimental results on four standard benchmarks including miniImageNet, tieredImageNet, CUB, and CIFAR-FS demonstrate superiority of our model in class-imbalanced generalization.