Toward A Logical Theory Of Fairness and Bias
This work addresses the issue of fairness in ML for researchers and practitioners by providing a formal grounding, but it is incremental as it builds on existing definitions without introducing a new paradigm.
The paper tackles the problem of formalizing fairness definitions in machine learning to address algorithmic bias, proposing a logical framework based on the epistemic situation calculus to reconstruct notions like fairness through unawareness, demographic parity, and counterfactual fairness.
Fairness in machine learning is of considerable interest in recent years owing to the propensity of algorithms trained on historical data to amplify and perpetuate historical biases. In this paper, we argue for a formal reconstruction of fairness definitions, not so much to replace existing definitions but to ground their application in an epistemic setting and allow for rich environmental modelling. Consequently we look into three notions: fairness through unawareness, demographic parity and counterfactual fairness, and formalise these in the epistemic situation calculus.