Reference-Limited Compositional Zero-Shot Learning

TL;DR

This paper introduces a new approach called MetaCGL for reference-limited compositional zero-shot learning, enabling recognition of unseen compositions with minimal reference data, and provides large-scale datasets for realistic evaluation.

Contribution

The paper proposes MetaCGL, a novel meta-learning graph-based model, and introduces new large-scale datasets for more realistic RL-CZSL evaluation.

Findings

MetaCGL achieves state-of-the-art accuracy in RL-CZSL tasks.

The datasets facilitate more realistic and challenging evaluations.

The method effectively generalizes from limited referential information.

Abstract

Compositional zero-shot learning (CZSL) refers to recognizing unseen compositions of known visual primitives, which is an essential ability for artificial intelligence systems to learn and understand the world. While considerable progress has been made on existing benchmarks, we suspect whether popular CZSL methods can address the challenges of few-shot and few referential compositions, which is common when learning in real-world unseen environments. To this end, we study the challenging reference-limited compositional zero-shot learning (RL-CZSL) problem in this paper, i.e., given limited seen compositions that contain only a few samples as reference, unseen compositions of observed primitives should be identified. We propose a novel Meta Compositional Graph Learner (MetaCGL) that can efficiently learn the compositionality from insufficient referential information and generalize to unseen compositions. Besides, we build a benchmark with two new large-scale datasets that consist of natural images with diverse compositional labels, providing more realistic environments for RL-CZSL. Extensive experiments in the benchmarks show that our method achieves state-of-the-art performance in recognizing unseen compositions when reference is limited for compositional learning.