AI Paper Atlas

Research map / Probabilistic and scientific modeling / Causal discovery

Relational Structural Causal Models

Adiba Ejaz, Elias Bareinboim

ICML 2026 · Probabilistic and scientific modeling › Causal discovery · approach: Counterfactual identification and bounds · 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
use relational neural causal model instead of flat NCM and REL-MLP baselinesRNCM · traffic scenes, identifiable source-target pairslog MSE of interventional query: often by about 100x lower errorImproved
train RNCM on a different source skeleton than the targetRNCM · traffic scenes, car c2 in target rho_Clog MSE of interventional query: outperforms NCM-X and NCM-J trained on targetImproved
train RNCM on source skeleton rho_A onlyRNCM · traffic scenes, car c1 (two signals)log MSE of interventional query: performance degradesHurt
combine rho_A and rho_B as training sources (ablation)RNCM · traffic scenes, car c1log MSE of interventional query: recovers performanceImproved
use degree-matched relational MLP (REL-MLP + DEG)REL-MLP · traffic scenes, car c3 without confoundinglog MSE of interventional query: strong on c3Mixed
run RelationalNeuralID on identifiable cross-car queryRNCM · unseen car traffic sceneslog MSE of interventional query: max-min gap collapses toward zeroImproved
run RelationalNeuralID on non-identifiable same-car queryRNCM · unseen car traffic scenesmax-min gap of query estimate: max-min gap remains largeNo 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