Invariant Causal Mechanisms through Distribution Matching
It addresses robustness in neural networks for domain generalization, but appears incremental as it builds on existing models.
The paper tackles the problem of learning invariant representations for data efficiency and robustness in neural networks, achieving state-of-the-art performance on domain generalization tasks with significant score boosts.
Learning representations that capture the underlying data generating process is a key problem for data efficient and robust use of neural networks. One key property for robustness which the learned representation should capture and which recently received a lot of attention is described by the notion of invariance. In this work we provide a causal perspective and new algorithm for learning invariant representations. Empirically we show that this algorithm works well on a diverse set of tasks and in particular we observe state-of-the-art performance on domain generalization, where we are able to significantly boost the score of existing models.