AI Paper Atlas

Research map / Probabilistic and scientific modeling / Causal discovery

Multi-task Learning of Order-Consistent Causal Graphs

Xinshi Chen, Haoran Sun, Caleb Ellington, Eric Xing, Le Song

NeurIPS 2021 · Probabilistic and scientific modeling › Causal discovery · approach: Differentiable and score-based DAG learning · 3 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
increase number of tasks K from 1 to 32 (MultiDAG)MultiDAG · Synthetic p=64, n=10SHD: 104 -> 66Improved
increase K from 1 to 32 (MultiDAG)MultiDAG · Synthetic p=64, n=320SHD: 19 -> 3Improved
increase K from 1 to 32 (JointGES)JointGES · Synthetic p=64, n=320SHD: 18 -> 26Hurt
increase K from 1 to 32 (JointGES)JointGES · Synthetic p=32, n=320SHD: 6 -> 9Hurt
increase K from 1 to 32 (MultiDAG)MultiDAG · Synthetic p=32, n=10SHD: 39 -> 23Improved
MultiDAG instead of JointGES at K=32MultiDAG · Synthetic p=64, n=320SHD: 26 -> 3Improved
MultiDAG instead of JointGES at K=1MultiDAG · Synthetic p=32, n=10SHD: 31 -> 39No clear effect
joint estimation K=10 versus single task K=1 at n=100MultiDAG · SERGIO E. coli 100SHD: 104.61 -> 84.0Improved
joint estimation K=10 versus single task K=1 at n=100MultiDAG · SERGIO E. coli 100FDR: 0.18 -> 0.07Improved
joint estimation K=10 versus single task K=1 at n=100MultiDAG · SERGIO E. coli 100TPR: 0.30 -> 0.42Improved
increase per-task samples n from 100 to 1000 at K=1MultiDAG · SERGIO E. coli 100SHD: 104.61 -> 76.4Improved
K=10,n=100 versus K=1,n=1000 (same total samples)MultiDAG · SERGIO E. coli 100SHD: 76.4 -> 84.0Mixed

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