Zero-Shot Recognition via Optimal Transport

TL;DR

This paper introduces an optimal transport framework for generalized zero-shot learning that aligns generated and real features to improve classification of seen and unseen classes, outperforming existing methods.

Contribution

It presents a novel OT-based approach combined with a generative model and attribute regularization for enhanced zero-shot learning performance.

Findings

Outperforms state-of-the-art on benchmark datasets

Effective in generalized, standard, and transductive ZSL

Improves discriminative power of generated features

Abstract

We propose an optimal transport (OT) framework for generalized zero-shot learning (GZSL), seeking to distinguish samples for both seen and unseen classes, with the assist of auxiliary attributes. The discrepancy between features and attributes is minimized by solving an optimal transport problem. {Specifically, we build a conditional generative model to generate features from seen-class attributes, and establish an optimal transport between the distribution of the generated features and that of the real features.} The generative model and the optimal transport are optimized iteratively with an attribute-based regularizer, that further enhances the discriminative power of the generated features. A classifier is learned based on the features generated for both the seen and unseen classes. In addition to generalized zero-shot learning, our framework is also applicable to standard and transductive ZSL problems. Experiments show that our optimal transport-based method outperforms state-of-the-art methods on several benchmark datasets.