Causal representation learning, causal generative AI, and beyond

As a core pillar of science and engineering, causality is transforming our approach to modern machine learning and artificial intelligence. Uncovering the causal process underlying observed data naturally helps answer 'why' and ‘what-if' questions, informs optimal decision-making, and enables adaptive prediction. In many scenarios, observed variables, such as image pixels and questionnaire responses, are often reflections of the underlying hidden causal variables rather than being causal variables themselves. Causal representation learning aims to reveal the underlying hidden causal variables and their relations. In this talk, we show how the modularity property of causal systems makes it possible to recover the underlying causal representations from observational data with identifiability guarantees. We further demonstrate how identifiable causal representation learning can directly benefit generative AI, using image generation / editing and extrapolative data generation as illustrative examples.