EVA: Mixture-of-Experts Semantic Variant Alignment for Compositional Zero-Shot Learning

TL;DR

EVA introduces a Mixture-of-Experts framework with semantic variant alignment to improve compositional zero-shot learning, achieving superior results by modeling high-quality primitives and fine-grained image-composition matching.

Contribution

The paper proposes EVA, a novel Mixture-of-Experts approach with semantic variant alignment for better primitive representation and compositional generalization in CZSL.

Findings

EVA outperforms state-of-the-art CZSL methods on three benchmarks.

The method improves primitive representation quality.

EVA enhances compositional generalization in both closed- and open-world settings.

Abstract

Compositional Zero-Shot Learning (CZSL) investigates compositional generalization capacity to recognize unknown state-object pairs based on learned primitive concepts. Existing CZSL methods typically derive primitives features through a simple composition-prototype mapping, which is suboptimal for a set of individuals that can be divided into distinct semantic subsets. Moreover, the all-to-one cross-modal primitives matching neglects compositional divergence within identical states or objects, limiting fine-grained image-composition alignment. In this study, we propose EVA, a Mixture-of-Experts Semantic Variant Alignment framework for CZSL. Specifically, we introduce domain-expert adaption, leveraging multiple experts to achieve token-aware learning and model high-quality primitive representations. To enable accurate compositional generalization, we further present semantic variant alignment to select semantically relevant representation for image-primitives matching. Our method significantly outperforms other state-of-the-art CZSL methods on three popular benchmarks in both closed- and open-world settings, demonstrating the efficacy of the proposed insight.