Learning soft interventions in complex equilibrium systems

TL;DR

This paper introduces a differentiable framework for soft interventions in cyclic causal models, enabling optimization of interventions in complex systems with feedback loops, exemplified through sustainable economy scenarios.

Contribution

It develops a novel Lie group-based approach for soft interventions and leverages automatic differentiation to optimize in cyclic causal systems, addressing a key challenge in complex system interventions.

Findings

Framework successfully models soft interventions in cyclic systems

Optimization of interventions improves transition scenarios to sustainability

Demonstrates applicability to economic transition models

Abstract

Complex systems often contain feedback loops that can be described as cyclic causal models. Intervening in such systems may lead to counterintuitive effects, which cannot be inferred directly from the graph structure. After establishing a framework for differentiable soft interventions based on Lie groups, we take advantage of modern automatic differentiation techniques and their application to implicit functions in order to optimize interventions in cyclic causal models. We illustrate the use of this framework by investigating scenarios of transition to sustainable economies.