Counterfactual Causality from First Principles?
This is an incremental position paper that critiques and suggests improvements for causality frameworks, primarily relevant to researchers in causal inference and AI.
The paper identifies three shortcomings in existing counterfactual causality approaches from a computer science perspective and proposes directions to address them, focusing on requirements-driven definitions, support for system dynamics, and behavior under abstraction.
In this position paper we discuss three main shortcomings of existing approaches to counterfactual causality from the computer science perspective, and sketch lines of work to try and overcome these issues: (1) causality definitions should be driven by a set of precisely specified requirements rather than specific examples; (2) causality frameworks should support system dynamics; (3) causality analysis should have a well-understood behavior in presence of abstraction.