Zero-Shot Learning posed as a Missing Data Problem

TL;DR

This paper introduces a novel zero-shot learning approach that models unseen class data distributions in the image feature space by leveraging label embeddings, demonstrating superior performance on benchmark datasets.

Contribution

It proposes a transductive framework that treats ZSL as a missing data problem, reversing traditional mappings to improve unseen class data estimation.

Findings

Outperforms state-of-the-art on two datasets

Effective transfer of knowledge from label embeddings

Reverses traditional ZSL mapping approach

Abstract

This paper presents a method of zero-shot learning (ZSL) which poses ZSL as the missing data problem, rather than the missing label problem. Specifically, most existing ZSL methods focus on learning mapping functions from the image feature space to the label embedding space. Whereas, the proposed method explores a simple yet effective transductive framework in the reverse way \--- our method estimates data distribution of unseen classes in the image feature space by transferring knowledge from the label embedding space. In experiments, our method outperforms the state-of-the-art on two popular datasets.