CAST (CausalSteward)

About

CAST (CausalSteward) is a research framework that combines LLM agents with divide-conquer-combine strategies to discover causal relationships in high-dimensional data. The system enables human-in-the-loop collaboration where AI agents help assemble and validate large causal models, making complex causal inference more tractable. Particularly useful for researchers working with observational data who need to build interpretable causal graphs at scale.

Details

Tags