1

We derive a minimal bound of experiments that guarantee identifiability of causal variables, opening up new opportunities for using intervention design for causal representation learning.

We present CITRIS, a causal representation learning algorithm for multidimensional causal factors identified from videos with interventions. Published at ICML 2022 (**Spotlight**).

We identify causal variables and their graphs from pairs of high-dimensional observations between which single interventions have been performed. Published at NeurIPS 2022.

We propose Mesh Convolution Neural Networks for rapid estimation of CFD parameters on arteries such as wall shear stress (WSS)

We present ENCO, an efficient structure learning method that leverages observational and interventional data and scales to graphs with a thousand variables. Published at ICLR 2022.

We explore the application of normalizing flows on categorical data, and propose GraphCNF a permutation-invariant generative model on graphs. Published at ICLR 2021.

We investigate deep-learning based heuristics for first-order automated theorem provers.