Towards Causal Representation Learning
It bridges machine learning and causality to tackle fundamental AI problems, but is incremental as it reviews and proposes rather than introduces new methods.
The paper reviews how causal inference concepts can address key machine learning challenges like transfer and generalization, and identifies causal representation learning as a central problem for discovering high-level causal variables from low-level observations.
The two fields of machine learning and graphical causality arose and developed separately. However, there is now cross-pollination and increasing interest in both fields to benefit from the advances of the other. In the present paper, we review fundamental concepts of causal inference and relate them to crucial open problems of machine learning, including transfer and generalization, thereby assaying how causality can contribute to modern machine learning research. This also applies in the opposite direction: we note that most work in causality starts from the premise that the causal variables are given. A central problem for AI and causality is, thus, causal representation learning, the discovery of high-level causal variables from low-level observations. Finally, we delineate some implications of causality for machine learning and propose key research areas at the intersection of both communities.