Towards Realistic Models of Change: A Causal Perspective on Decision-Making and Dynamical Systems

Key points

• While statistical learning has made remarkable progress in recent decades, driven by large datasets and data-driven protocols that have transformed scientific practice, it still faces a key limitation: it struggles to generalize beyond the data it has seen, often failing under distribution shift.1
• We propose an alternative perspective to the established framework, arguing for a more realistic cost function, a reformulation of the problem in causal terms, and the inclusion of a temporal dimension in policy evaluation.1
• The second part focuses on dynamical systems, extending the notions of causal interventions and effects to discrete and continuous stochastic processes.1
• We show that the equilibria of these processes, and how they change following an intervention, can be linked to Structural Causal Models (SCMs), viewed as compressed representations of the underlying temporal dynamics.1

Sources (1)

• [1]Towards Realistic Models of Change: A Causal Perspective on Decision-Making and Dynamical Systems
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 20, 06:26 PM

  Causality offers a promising solution: causal models focus on the underlying mechanisms of change rather than surface-level associations, enabling robust reasoning under external interventions.

  While statistical learning has made remarkable progress in recent decades, driven by large datasets and data-driven protocols that have transformed scientific practice, it still faces a key limitation: it struggles to generalize beyond the data it has seen, often failing under distribution shift.

Extractive summary: sentences quoted from the sources.