Learning Primitive Relations for Compositional Zero-Shot Learning

TL;DR

This paper introduces a novel framework called learning primitive relations (LPR) for compositional zero-shot learning, which models state-object relationships using cross-attention to improve unseen composition prediction.

Contribution

The paper proposes a probabilistic framework with cross-attention to explicitly model state-object dependencies, advancing CZSL performance.

Findings

LPR outperforms state-of-the-art methods on benchmark datasets.

LPR effectively leverages state-object relationships for unseen compositions.

The approach improves both closed-world and open-world CZSL settings.

Abstract

Compositional Zero-Shot Learning (CZSL) aims to identify unseen state-object compositions by leveraging knowledge learned from seen compositions. Existing approaches often independently predict states and objects, overlooking their relationships. In this paper, we propose a novel framework, learning primitive relations (LPR), designed to probabilistically capture the relationships between states and objects. By employing the cross-attention mechanism, LPR considers the dependencies between states and objects, enabling the model to infer the likelihood of unseen compositions. Experimental results demonstrate that LPR outperforms state-of-the-art methods on all three CZSL benchmark datasets in both closed-world and open-world settings. Through qualitative analysis, we show that LPR leverages state-object relationships for unseen composition prediction.