A Generative Adversarial Approach for Zero-Shot Learning from Noisy Texts

TL;DR

This paper introduces a GAN-based method that generates visual features from noisy text descriptions to improve zero-shot learning, outperforming previous approaches on large benchmarks.

Contribution

It proposes a simple generative model with visual pivot regularization to effectively handle noisy text data in zero-shot learning.

Findings

Outperforms state-of-the-art on large text-based zero-shot learning benchmarks.

Effectively suppresses noise without complex regularizers.

Transforms zero-shot learning into a standard classification task.

Abstract

Most existing zero-shot learning methods consider the problem as a visual semantic embedding one. Given the demonstrated capability of Generative Adversarial Networks(GANs) to generate images, we instead leverage GANs to imagine unseen categories from text descriptions and hence recognize novel classes with no examples being seen. Specifically, we propose a simple yet effective generative model that takes as input noisy text descriptions about an unseen class (e.g.Wikipedia articles) and generates synthesized visual features for this class. With added pseudo data, zero-shot learning is naturally converted to a traditional classification problem. Additionally, to preserve the inter-class discrimination of the generated features, a visual pivot regularization is proposed as an explicit supervision. Unlike previous methods using complex engineered regularizers, our approach can suppress the noise well without additional regularization. Empirically, we show that our method consistently outperforms the state of the art on the largest available benchmarks on Text-based Zero-shot Learning.