Research
158d agoMARLIN Accelerates Causal Discovery
The MARLIN framework leverages multi‑agent RL to uncover causal directed acyclic graphs from observational data more efficiently than existing methods. By mapping continuous spaces to DAGs and integrating state‑specific and state‑invariant agents, MARLIN enables incremental learning suitable for real‑time applications.
The Signal
Its scalable approach promises faster causal inference across scientific, economic, and policy domains worldwide.