Learning soft interventions in complex equilibrium systems
This work addresses the challenge of predicting and controlling counterintuitive effects in complex systems, such as economic transitions, for researchers and practitioners in causal inference and systems modeling.
The paper tackled the problem of optimizing interventions in complex equilibrium systems with feedback loops by developing a differentiable soft intervention framework based on Lie groups, leveraging automatic differentiation to handle cyclic causal models, and demonstrated its application in scenarios for transitioning to sustainable economies.
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.