Statistics

Statistical Methods for xAI

Explainable Artificial Intelligence requires appropriate statistical metrics to assess explainability and, more generally, trustworthiness of machine learning output. This leads to SAFE machine learning (Sustainability, Accuracy, Fairness, Explainability) in which AI can measure its own risks and become more responsible. 

Following the organisation of conferences and satellite meetings on Explainable AI and on SAFE machine learning, the journal is excited to announce a special issue focused on Statistical Methods for Explainable Artificial Intelligence. This special issue can include the extended versions of the papers selected at the conferences and, more generally, all selected papers that include statistical approaches for explainable and SAFE artificial intelligence. Topics for the special issue include, but are not limited to the following:

• Statistical measures of accuracy
• Statistical measures of sustainability
• Statistical measures of explainability
• Statistical measures of fairness
• Model regularisation to improve explainability
• Dimension reduction methods to improve explainability
• Improving robustness of explanations
• Sensitivity analysis for xAI methods
• Explainable methods and fairness
• Group based fairness for xAi
• Counterfactual fairness with/for xAI methods
• Statistical tests for accuracy
• Statistical tests for explainability
• Statistical tests for fairness
• Statistical tests for robustness