An Iterative Co-Training Transductive Framework for Zero Shot Learning

TL;DR

This paper introduces an iterative co-training framework for zero-shot learning that leverages two models exchanging pseudo labels to improve unseen-class classification, also addressing bias in generalized ZSL.

Contribution

It proposes a novel co-training approach with an exchanging module for pseudo labels and a semantic-guided OOD detector for GZSL, enhancing performance over existing methods.

Findings

Significantly outperforms about 31 state-of-the-art methods

Effectively exploits model complementarity through iterative pseudo label exchange

Reduces bias in generalized ZSL with semantic-guided OOD detection

Abstract

In zero-shot learning (ZSL) community, it is generally recognized that transductive learning performs better than inductive one as the unseen-class samples are also used in its training stage. How to generate pseudo labels for unseen-class samples and how to use such usually noisy pseudo labels are two critical issues in transductive learning. In this work, we introduce an iterative co-training framework which contains two different base ZSL models and an exchanging module. At each iteration, the two different ZSL models are co-trained to separately predict pseudo labels for the unseen-class samples, and the exchanging module exchanges the predicted pseudo labels, then the exchanged pseudo-labeled samples are added into the training sets for the next iteration. By such, our framework can gradually boost the ZSL performance by fully exploiting the potential complementarity of the two models' classification capabilities. In addition, our co-training framework is also applied to the generalized ZSL (GZSL), in which a semantic-guided OOD detector is proposed to pick out the most likely unseen-class samples before class-level classification to alleviate the bias problem in GZSL. Extensive experiments on three benchmarks show that our proposed methods could significantly outperform about $31$ state-of-the-art ones.