Learning Invariant Visual Representations for Compositional Zero-Shot Learning
This work addresses the challenge of recognizing novel attribute-object compositions in computer vision, which is crucial for applications like robotics and image understanding, though it appears incremental as it builds on existing CZSL methods by incorporating invariance techniques.
The paper tackles the problem of Compositional Zero-Shot Learning (CZSL) by reframing it as an out-of-distribution generalization issue, proposing an invariant feature learning framework that aligns domains at representation and gradient levels to reduce spurious correlations. Experiments on two benchmarks show the method significantly outperforms previous state-of-the-art approaches.
Compositional Zero-Shot Learning (CZSL) aims to recognize novel compositions using knowledge learned from seen attribute-object compositions in the training set. Previous works mainly project an image and a composition into a common embedding space to measure their compatibility score. However, both attributes and objects share the visual representations learned above, leading the model to exploit spurious correlations and bias towards seen pairs. Instead, we reconsider CZSL as an out-of-distribution generalization problem. If an object is treated as a domain, we can learn object-invariant features to recognize the attributes attached to any object reliably. Similarly, attribute-invariant features can also be learned when recognizing the objects with attributes as domains. Specifically, we propose an invariant feature learning framework to align different domains at the representation and gradient levels to capture the intrinsic characteristics associated with the tasks. Experiments on two CZSL benchmarks demonstrate that the proposed method significantly outperforms the previous state-of-the-art.