Causal Inference With Deep Generative Models

Causal knowledge is central to solving complex decision-making problems in many fields, from engineering and medicine to cyber-physical systems. Causal inference has also recently been identified as a key capability to remedy some of the issues modern machine learning systems suffer from, such as explainability and generalization.

In this talk, we first provide a short introduction to the principles of causal modeling. Next, we discuss how deep neural networks can be used to obtain a causal representation to solve complex high-dimensional and distributed causal inference problems using deep generative models. Finally, we show how the proposed methods can be applied for causal invariant prediction and to evaluate black-box conditional generative models.

Murat Kocaoglu is currently an assistant professor in the Department of Computer Science at Johns Hopkins University, where he leads the CausalML Lab. His current research interests include causal inference, deep causal generative models, and information theory.

Organized by:

Institute of AI in Management
LMU Munich