Causal analysis, Correlation-Response and Dynamic cavity

TL;DR

This paper explores the analogies between causal analysis in statistics and correlation-response theory in physics, proposing that dynamic cavity methods can effectively compute stationary states in non-equilibrium systems, offering new perspectives for causal analysis.

Contribution

It introduces the connection between causal analysis and correlation-response theory and suggests dynamic cavity as a tool for analyzing non-equilibrium stationary states.

Findings

Analogies between causal analysis and correlation-response theory identified.

Dynamic cavity method can compute stationary states in non-equilibrium systems.

Proposes an alternative approach to causal analysis using physics-inspired methods.

Abstract

The purpose of this note is to point out analogies between causal analysis in statistics and the correlation-response theory in statistical physics. It is further shown that for some systems the dynamic cavity offers a way to compute the stationary state of a non- equilibrium process effectively, which could then be taken an alternative starting point of causal analysis.