arXiv Machine Learning
Differentiable Causal Discovery for Singular Linear Models under Confounding
Read the original on arXiv Machine Learning →The Flow has not summarised this story yet — read it at arXiv Machine Learning.
Related stories
arXiv AI
Jul 13
How Does Bayesian Causal Discovery Fail? Characterising Structural Consequences in Linear Gaussian Networks under Latent Confounding
arXiv:2607. 09449v1 Announce Type: new Abstract: Bayesian causal discovery is widely used for its ability to quantify epistemic uncertainty over directed acyclic graphs (DAGs) through posterior inference.
arXiv AI
Oct 21
Causal Discovery for Linear Models with Latent Confounding via Multi-domain Identifiable Representation Learning
arXiv:2610.28134v1 Announce Type: cross Abstract: Causal discovery with latent confounding is particularly challenging, but a common setting in real-world data. In this work, we focus on linear causa...
arXiv AI
Aug 6
SVI-DAG: A Structured Variational Inference Approach to Bayesian Causal Discovery
arXiv:2608. 04930v1 Announce Type: cross Abstract: Bayesian causal discovery seeks to determine the posterior distribution of causal theories, which are interpreted as directed acyclic graphs (DAGs) that explain the observed data.
arXiv Statistics ML
Sep 28
MARCEDES: Score-based causal discovery under non-Gaussianity with continuous optimization
arXiv:2609.30643v1 Announce Type: new Abstract: We consider the problem of learning the underlying causal directed acyclic graph (DAG) structure corresponding to a structural equation model (SEM) wit...
arXiv AI
Oct 1
Differentiable Structure Learning for Cyclic Linear Gaussian Models with Latent Confounders
The paper introduces a method for learning causal structures in linear Gaussian models that may contain directed cycles and an unknown number of latent confounders, bounded by a maximum. It derives the covariance of observed variables, defines marginal quasi-equivalence to identify when different models produce the same observational distributions, and formulates structure learning as a minimization of Gaussian negative log-likelihood with a complexity penalty counting edges and latent variables. Using Bernoulli gates to parameterize edge and latent inclusion, the authors obtain a closed‑form differentiable objective whose expected value shares the same global optimum as the discrete problem, and demonstrate experimentally that this approach yields lower recovery error than prior methods.
arXiv Statistics ML
6d ago
Multi-Domain Empirical Bayes for Linearly-Mixed Causal Representations
arXiv:2603.18404v3 Announce Type: replace Abstract: Causal representation learning (CRL) aims to learn low-dimensional causal latent variables from high-dimensional observations. While identifiabilit...