AI Paper Atlas

Research map / Probabilistic and scientific modeling / Causal discovery

One Intervention per Component is Enough: Towards Identifiability in Linear Stochastic Dynamics from Steady State

Saber Salehkaleybar

ICML 2026 · Probabilistic and scientific modeling › Causal discovery · approach: Latent-variable causal estimation · 0 citations (OpenAlex)

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
add mean-equation residuals to the covariance-only loss (ablation)Regularized least-squares OU estimator · Synthetic OU, n=10relative error of Lambda, D, bImproved
add mean-equation residuals to covariance-only loss (ablation)OU estimator on Perturb-seq · Perturb-seq Co-CulturePDS: 0.43 -> 0.67Improved
add mean-equation residuals to covariance-only loss (ablation)OU estimator on Perturb-seq · Perturb-seq Co-CultureDES: 0.51 -> 0.43Hurt
add mean-equation residuals to covariance-only loss (ablation)OU estimator on Perturb-seq · Perturb-seq ControlDES: 0.46 -> 0.33Hurt
add mean-equation residuals to covariance-only loss (ablation)OU estimator on Perturb-seq · Perturb-seq IFN-gammaPDS: 0.43 -> 0.70Improved
add mean-equation residuals to covariance-only loss (ablation)OU estimator on Perturb-seq · Perturb-seq IFN-gammaDES: 0.42 -> 0.47Improved
increase the number of interventions from 0 to 10Regularized least-squares OU estimator · Non-diagonal diffusion syntheticrelative error of Lambda: 0.2064 -> 0.0646Improved
increase the number of interventions from 0 to 10Regularized least-squares OU estimator · Non-diagonal diffusion syntheticrelative error of D: 0.1811 -> 0.0719Improved
increase the number of interventions from 0 to 10Regularized least-squares OU estimator · Non-diagonal diffusion syntheticrelative error of b: 0.1407 -> 0.0661Improved
increase the number of interventions on diagonal-diffusion D estimateRegularized least-squares OU estimator · Synthetic OU, n=10relative error of DMixed
increase sample size from 500 to 10000 for hard interventionsSCC recovery from mean shifts · Synthetic SCC recoveryAdjusted Rand Index: 0.660 -> 0.881Improved
increase sample size from 500 to 10000 for soft interventionsSCC recovery from mean shifts · Synthetic SCC recoveryAdjusted Rand Index: 0.464 -> 0.837Improved

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