J9VogDTa1W@OpenReview

Total: 1

#1 Systems with Switching Causal Relations: A Meta-Causal Perspective [PDF6] [Copy] [Kimi13] [REL]

Authors: Moritz Willig, Tim Tobiasch, Florian Busch, Jonas Seng, Devendra Singh Dhami, Kristian Kersting

Most works on causality in machine learning assume that causal relationships are governed by a constant underlying process. However, the flexibility of agents' actions or tipping point behavior in the environmental process can change the qualitative dynamics of the system. As a result, new causal relationships may emerge, while existing ones change or disappear, resulting in an altered causal graph. To analyze these qualitative changes on the causal graph, we propose the concept of *meta-causal states*, which groups classical causal models into clusters based on equivalent qualitative behavior and consolidates specific mechanism parameterizations. We demonstrate how meta-causal states can be inferred from observed agent behavior, and discuss potential methods for disentangling these states from unlabeled data. Finally, we direct our analysis toward the application of a dynamical system, demonstrating that meta-causal states can also emerge from inherent system dynamics, and thus constitute more than a context-dependent framework in which mechanisms emerge merely as a result of external factors.

Subject: ICLR.2025 - Spotlight