Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Test Time Training for Supervised Causal Learning

About

Supervised Causal Learning (SCL) has shown promise in causal discovery by framing it as a supervised learning problem. However, it suffers from significant out-of-distribution generalization challenges. We reveal three limitations of previous SCL practices: a significant performance gap between synthetic benchmarks and real-world data, fragility to distribution shifts, and failure in compositional generalization, collectively questioning its real-world applicability. To address this, we propose Test-Time Training for Supervised Causal Learning (TTT-SCL), a novel framework that dynamically generates training sets explicitly aligned with any specific test instance. We demonstrate the correlation between TTT-SCL and score-based methods, and design an efficient module for generating training sets based on the classic scoring function. Experiments on synthetic benchmarks, pseudo-real and real-world datasets demonstrate that TTT-SCL significantly outperforms existing SCL and traditional causal discovery methods.

Zizhen Deng, Jiaru Zhang, Rui Ding, Huang Bojun, Jinzhuo Wang, Qiang Fu, Shi Han, Dongmei Zhang• 2026

Related benchmarks

TaskDatasetResultRank
Causal DiscoverySyntren
F1 Score32.14
22
Edge PredictionLinear_U
AUROC86.3
9
Edge PredictionChebyshev_G
AUROC83
9
Edge PredictionSachs
AUROC78.9
9
Edge PredictionSyntren
AUROC80.1
9
Edge PredictionRFF_G
AUROC91.8
9
Causal DiscoveryAsia bnlearn
AUROC91
6
Causal DiscoveryCancer bnlearn
AUROC91.6
6
Causal DiscoveryEarthquake bnlearn repository
AUROC98.8
6
Causal DiscoverySurvey bnlearn
AUROC95.5
6
Showing 10 of 14 rows

Other info

Follow for update