Towards Zero-Shot Learning with Fewer Seen Class Examples

TL;DR

This paper introduces a meta-learning based generative model for zero-shot learning that effectively handles scenarios with very few seen class examples, outperforming existing methods on benchmark datasets.

Contribution

It proposes a novel meta-learning framework combining VAE and GAN to generate high-quality samples from limited seen class data for improved ZSL.

Findings

Outperforms state-of-the-art methods with few seen examples

Effective in generating high-fidelity samples from limited data

Demonstrates robustness across multiple benchmark datasets

Abstract

We present a meta-learning based generative model for zero-shot learning (ZSL) towards a challenging setting when the number of training examples from each \emph{seen} class is very few. This setup contrasts with the conventional ZSL approaches, where training typically assumes the availability of a sufficiently large number of training examples from each of the seen classes. The proposed approach leverages meta-learning to train a deep generative model that integrates variational autoencoder and generative adversarial networks. We propose a novel task distribution where meta-train and meta-validation classes are disjoint to simulate the ZSL behaviour in training. Once trained, the model can generate synthetic examples from seen and unseen classes. Synthesize samples can then be used to train the ZSL framework in a supervised manner. The meta-learner enables our model to generates high-fidelity samples using only a small number of training examples from seen classes. We conduct extensive experiments and ablation studies on four benchmark datasets of ZSL and observe that the proposed model outperforms state-of-the-art approaches by a significant margin when the number of examples per seen class is very small.