MIT News November 1, 2024
Despite its central role, current methods for causal inference face significant challenges due to nonlinear dependencies, stochastic interactions, self-causation, collider effects, and influences from exogenous factors, among others. While existing methods can effectively address some of these challenges, no single approach has successfully integrated all these aspects. A team of researchers in the US (MIT, Caltech) developed SURD: Synergistic-Unique-Redundant Decomposition of causality which quantified causality as the increments of redundant, unique, and synergistic information gained about future events from past observations. The formulation was non-intrusive and applicable to both computational and experimental investigations, even when samples were scarce. They benchmarked SURD in scenarios that posed significant challenges for causal inference and demonstrated that it offered a more reliable quantification of causality compared to previous methods… read more. Open Access TECHNICAL ARTICLEÂ
Schematic of the causal decomposition in SURD. Credit: Nature Communications volume 15, Article number: 9296, 1 November 2024Â