Showing 1–1 of 1 results for author: Privato, N

A Context-Sensitive Approach to XAI in Music Performance

Authors: Nicola Privato, Jack Armitage

Abstract: The rapidly evolving field of Explainable Artificial Intelligence (XAI) has generated significant interest in develo** methods to make AI systems more transparent and understandable. However, the problem of explainability cannot be exhaustively solved in the abstract, as there is no single approach that can be universally applied to generate adequate explanations for any given AI system, and thi… ▽ More The rapidly evolving field of Explainable Artificial Intelligence (XAI) has generated significant interest in develo** methods to make AI systems more transparent and understandable. However, the problem of explainability cannot be exhaustively solved in the abstract, as there is no single approach that can be universally applied to generate adequate explanations for any given AI system, and this is especially true in the arts. In this position paper, we propose an Explanatory Pragmatism (EP) framework for XAI in music performance, emphasising the importance of context and audience in the development of explainability requirements. By tailoring explanations to specific audiences and continuously refining them based on feedback, EP offers a promising direction for enhancing the transparency and interpretability of AI systems in broad artistic applications and more specifically to music performance. △ Less

Submitted 5 September, 2023; originally announced September 2023.