Generative Interventions for Causal Learning

Abstract

We introduce a framework for learning robust visual representations that generalize to new viewpoints, backgrounds, and scene contexts. Discriminative models often learn naturally occurring spurious correlations, which cause them to fail on images outside of the training distribution. In this paper, we show that we can steer generative models to manufacture interventions on features caused by confounding factors. Experiments, visualizations, and theoretical results show this method learns robust representations more consistent with the underlying causal relationships. Our approach improves performance on multiple datasets demanding out-of-distribution generalization, and we demonstrate state-of-the-art performance generalizing from ImageNet to ObjectNet dataset.

Accepted to CVPR 2021

References in corpus (6)

• Improved Regularization of Convolutional Neural Networks with Cutout
• The Effectiveness of Data Augmentation in Image Classification using Deep Learning
• CXPlain: Causal Explanations for Model Interpretation under Uncertainty
• Causal Induction from Visual Observations for Goal Directed Tasks
• Continuously Indexed Domain Adaptation
• Deep causal representation learning for unsupervised domain adaptation