AI Paper Atlas

Research map / Graph learning / Graph algorithms and clustering

On Learning Necessary and Sufficient Causal Graphs

Hengrui Cai, Yixin Wang, Michael Jordan, Rui Song

NeurIPS 2023 · Graph learning › Graph algorithms and clustering · approach: Continuous DAG structure learning · 1 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 total effect instead of direct effect as criterion (ablation)NSCSL · S2 p=5, n=100TPR: 0.50 -> 0.99Improved
use total effect instead of direct effect as criterion (ablation)NSCSL · S4 ER p=20, n=1000TPR: 0.67 -> 1.00Improved
use total effect instead of direct effect as criterion (ablation)NSCSL · S5 ER p=50, n=3000SHD: 9.76 -> 1.58Improved
use total effect instead of direct effect as criterion (ablation)NSCSL · S1 p=5, n=100SHD: 0.14 -> 0.14No clear effect
add necessary-and-sufficient causal effect penalty to NOTEARSNOTEARS · S1 p=5, n=100SHD: 2.02 -> 0.14Improved
add necessary-and-sufficient causal effect penalty to NOTEARSNOTEARS · S4 ER p=20, n=1000SHD: 36.40 -> 0.04Improved
add causal effect penalty to NOTEARS in a setting with no spurious variablesNOTEARS · S3 p=5, n=100SHD: 0.06 -> 0.06No clear effect
add causal effect penalty to NOTEARS under nonlinear SEMNOTEARS · S4 nonlinear p=20, n=1000SHD: 35.90 -> 0.60Improved
add causal effect penalty to NOTEARS under scale-free graphNOTEARS · S5 SF p=50, n=1000SHD: 123.10 -> 5.08Improved
increase sample size from 30 to 100NSCSL-TE · S1 p=5SHD: 0.64 -> 0.14Improved
increase sample size from 100 to 1000NSCSL-TE · S4 ER p=20runtime: 10.8 -> 56.2Hurt
add causal effect penalty to NOTEARS on single-cell protein dataNOTEARS · Akt protein networkSHD: 21 -> 8Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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