ResearchResearch paperInterpretability · Training & Scaling · Reinforcement Learning1 source · Oct 21, 2026

SCDM: Scalable Causal Discovery in Nonlinear Temporal Systems with Meta-Learning

We propose SCDM, a shared-parameter meta-learning framework for scalable temporal causal discovery.

Causal discovery from nonlinear multivariate time series is challenging in high-dimensional systems, where the number of candidate directed relations grows quadratically with the number of variables.
SCDM treats each target variable as a task while learning a shared temporal predictor and a shared causal-strength matrix across tasks.
After training, a post-training graph readout converts the learned predictor into continuous directed causal scores for threshold-independent evaluation.
Experiments on controlled synthetic, neuroimaging-inspired, high-dimensional, and real-world temporal benchmarks demonstrate competitive causal recovery and improved scalability in large temporal systems.

[1]SCDM: Scalable Causal Discovery in Nonlinear Temporal Systems with Meta-Learning
arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 21, 04:31 PM
We propose SCDM, a shared-parameter meta-learning framework for scalable temporal causal discovery.
Causal discovery from nonlinear multivariate time series is challenging in high-dimensional systems, where the number of candidate directed relations grows quadratically with the number of variables.

Extractive summary: sentences quoted from the sources.

Before this