Heterogeneous Graph-based Knowledge Transfer for Generalized Zero-shot Learning

TL;DR

This paper introduces a heterogeneous graph neural network approach for generalized zero-shot learning that effectively transfers knowledge from seen to unseen classes without prior information about the unseen classes, achieving state-of-the-art results.

Contribution

The paper proposes a novel heterogeneous graph-based method that is agnostic to unseen classes, using Wasserstein barycenter for class representation and GNN for knowledge transfer.

Findings

Achieves state-of-the-art results on benchmark datasets.

Effectively transfers knowledge without prior unseen class information.

Utilizes Wasserstein barycenter for class representation.

Abstract

Generalized zero-shot learning (GZSL) tackles the problem of learning to classify instances involving both seen classes and unseen ones. The key issue is how to effectively transfer the model learned from seen classes to unseen classes. Existing works in GZSL usually assume that some prior information about unseen classes are available. However, such an assumption is unrealistic when new unseen classes appear dynamically. To this end, we propose a novel heterogeneous graph-based knowledge transfer method (HGKT) for GZSL, agnostic to unseen classes and instances, by leveraging graph neural network. Specifically, a structured heterogeneous graph is constructed with high-level representative nodes for seen classes, which are chosen through Wasserstein barycenter in order to simultaneously capture inter-class and intra-class relationship. The aggregation and embedding functions can be learned through graph neural network, which can be used to compute the embeddings of unseen classes by transferring the knowledge from their neighbors. Extensive experiments on public benchmark datasets show that our method achieves state-of-the-art results.