AI Paper Atlas

Research map / Graph learning / Graph algorithms and clustering

CauScale: Neural Causal Discovery at Scale

Bo Peng, Sirui Chen, Jiaguo Tian, Yu Qiao, Chaochao Lu

ICML 2026 · Graph learning › Graph algorithms and clustering · approach: Continuous DAG structure learning · 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 the reduction unit that average-pools data embeddings along the observation axis (ablation)CauScale · synthetic, n up to 100mAP, inference time, GPU memoryImproved
use average pooling instead of max pooling in the reduction unit (ablation)CauScale reduction unit · synthetic, n=50mAP: 71.2 -> 85.4Improved
use average pooling instead of strided pooling in the reduction unit (ablation)CauScale reduction unit · synthetic, n=50mAP: 78.5 -> 85.4Improved
add the graph stream alongside the data stream (ablation)CauScale · synthetic benchmarkmAPImproved
add the inverse-covariance graph prior as graph input (ablation)CauScale · synthetic benchmarkmAPImproved
use tied attention weights instead of vanilla attention (ablation)CauScale · synthetic, n=100, |E|=400mAP and inference time: mAP 0.67 -> 0.81; inference time 0.108 -> 0.038 sImproved
use the pairwise decomposed prediction head instead of the vanilla AVICI head (ablation)CauScale · synthetic, n up to 100cyclicity: 0.0-0.25% -> 0.0%Improved
use the pairwise decomposed prediction head instead of the vanilla AVICI head (ablation)CauScale · synthetic, n=100, |E|=400mAP: 0.81 vs 0.81No clear effect
apply cycle-breaking post-processing to predicted graphs (ablation)CauScale at n=1000 · synthetic Polynomial, n=1000SHD: 3985.0 -> 3891.8Improved
estimate graph prior from fewer samples (ablation)CauScale · n=1000, |E|=2000mAP: 64.9 (500 samples) -> 32.6 (10 samples)Hurt
hide high-out-degree nodes as latent confounders (ablation)CauScale · n=100, linearmAPMixed
test on OOD noise and mechanism functionsCauScale · OOD generalizationmAPMixed

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