Discrete Causal Representation Learning

Add to your list(s) Download to your calendar using vCal

• Yuqi Gu (Columbia University)
• Thursday 05 March 2026, 14:00-14:45
• Seminar Room 1, Newton Institute.

If you have a question about this talk, please contact nobody.

CIFW02 - Causal identification and discovery

Causal representation learning seeks to uncover causal relationships among high-level latent variables from low-level, entangled, and noisy observations. Existing approaches often either rely on deep neural networks, which lack interpretability and formal guarantees, or impose restrictive assumptions like linearity, continuous-only observations, and strong structural priors. These limitations particularly challenge applications with a large number of discrete latent variables and mixed-type observations. To address these challenges, we propose discrete causal representation learning, a generative framework that models a directed acyclic graph among discrete latent variables, along with a sparse bipartite graph linking latent and observed layers. This design accommodates continuous, count, and binary responses through flexible measurement models while maintaining interpretability. Under mild conditions, we prove both the bipartite measurement graph and the latent causal graph are identifiable. We further propose a three-stage estimate-resample-discovery pipeline: penalized estimation of the generative model parameters, resampling of latent configurations from the fitted model, and score-based causal discovery on the resampled latents. We establish the consistency of this procedure, ensuring reliable latent causal structure recovery. Empirical studies on educational assessment and synthetic image data demonstrate that discrete causal representation learning recovers sparse and interpretable latent causal structures.

This talk is part of the Isaac Newton Institute Seminar Series series.

This talk is included in these lists:

• All CMS events
• Featured lists
• INI info aggregator
• Isaac Newton Institute Seminar Series
• School of Physical Sciences
• Seminar Room 1, Newton Institute
• bld31
• dh539

Note that ex-directory lists are not shown.