audit volatility

New geometric framework analyzes robustness of AI fairness audits

Researchers have developed a new geometric framework to analyze the robustness of neighborhood-based fairness audits in machine learning. These audits assess individual fairness by comparing predictions for similar indi…