A causal view of compositional zero-shot recognition
This addresses the challenge of recognizing new visual categories composed of known components, which is critical for real-world domains like vision and language, but is incremental as it builds on existing causal ideas.
The paper tackles the problem of compositional generalization in zero-shot recognition by proposing a causal-inspired embedding model that learns disentangled representations from correlated training data, showing improvements on synthesized and real-world datasets compared to strong baselines.
People easily recognize new visual categories that are new combinations of known components. This compositional generalization capacity is critical for learning in real-world domains like vision and language because the long tail of new combinations dominates the distribution. Unfortunately, learning systems struggle with compositional generalization because they often build on features that are correlated with class labels even if they are not "essential" for the class. This leads to consistent misclassification of samples from a new distribution, like new combinations of known components. Here we describe an approach for compositional generalization that builds on causal ideas. First, we describe compositional zero-shot learning from a causal perspective, and propose to view zero-shot inference as finding "which intervention caused the image?". Second, we present a causal-inspired embedding model that learns disentangled representations of elementary components of visual objects from correlated (confounded) training data. We evaluate this approach on two datasets for predicting new combinations of attribute-object pairs: A well-controlled synthesized images dataset and a real-world dataset which consists of fine-grained types of shoes. We show improvements compared to strong baselines.