AI Paper Atlas

Research map / Probabilistic and scientific modeling / Causal discovery

Identifying dependent components from multi-domain linear mixtures

Danru Xu, Lauri Parkkonen, Sara Magliacane, Aapo Hyvarinen

ICML 2026 · Probabilistic and scientific modeling › Causal discovery · approach: Latent-variable causal estimation

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use MuDo-CoM instead of FastICA on correlated latentsFastICA · Simulated, ER-2 Gaussian SCM, n=10d-MCC: 0.783 -> 0.999Improved
use MuDo-CoM instead of QNDIAG on correlated latentsQNDIAG · Simulated, ER-2 Gaussian SCM, n=10d-MCC: 0.823 -> 0.999Improved
increase latent dependency degree from ER-0 to ER-3FastICA · Simulated, n=10d-MCC: 1.000 -> 0.745Hurt
increase latent dependency degree from ER-0 to ER-3MuDo-CoM · Simulated, n=10d-MCC: 0.994 -> 0.999No clear effect
use MuDo-CoM when latents are independent (ER-0)FastICA and QNDIAG · Simulated, ER-0d-MCC: 1.000 vs 0.994Hurt
increase latent dimension n from 5 to 20PARAFAC · Simulated, ER-2d-MCC: 0.688 -> 0.319Hurt
increase latent dimension n from 5 to 20MuDo-CoM · Simulated, ER-2d-MCC: 0.999 -> 0.998No clear effect
increase sample size N from 200 to 10000MuDo-CoM · Simulated, n=10d-MCC: 0.900 -> 1.000Improved
increase number of domains T from 1n to 3nMuDo-CoM · Simulated, n=10d-MCC: 0.998 -> 0.991No clear effect
increase observed dimension d from n to 10nMuDo-CoM · Simulated, n=10d-MCC: 0.999 -> 0.999No clear effect
use exponential or Gumbel noise instead of GaussianMuDo-CoM · Simulated, n=10d-MCC: 0.999 -> 0.991 (Exp), 0.996 (Gumbel)No clear effect
use nonlinear SCM instead of linearMuDo-CoM · Simulated, n=10d-MCC: 0.999 -> 0.999No clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy