Timezone: »

 
Latent Variable Models for Bayesian Causal Discovery
Jithendaraa Subramanian · Jithendaraa Subramanian · Yashas Annadani · Ivaxi Sheth · Stefan Bauer · Derek Nowrouzezahrai · Samira Ebrahimi Kahou

@
in Spurious correlations, Invariance, and Stability (SCIS) »
Event URL: https://openreview.net/forum?id=Au60kZskzgH »

Learning predictors that do not rely on spurious correlations involves building causal representations. However, learning such a representation is very challenging. We, therefore, formulate the problem of learning a causal representation from high dimensional data and study causal recovery with synthetic data. This work introduces a latent variable decoder model, Decoder BCD, for Bayesian causal discovery and performs experiments in mildly supervised and unsupervised settings. We present a series of synthetic experiments to characterize important factors for causal discovery.

Keywords: Latent Variable models · Causal Representation Learning · Bayesian Inference

Author Information

Jithendaraa Subramanian (McGill University, Mila)
Jithendaraa Subramanian

Masters student at McGill interested in building causal representations of the world into deep learning systems

Jithendaraa Subramanian (McGill University)
Yashas Annadani (ETH Zurich)
Ivaxi Sheth (Montreal Institute for Learning Algorithms, University of Montreal, Université de Montréal)
Stefan Bauer (KTH Stockholm)
Derek Nowrouzezahrai (McGill University)
Samira Ebrahimi Kahou (Microsoft Research)

More from the Same Authors