MUSE-Explainer: Counterfactual Explanations for Symbolic Music Graph Classification Models
This addresses the need for human-friendly explanations in music analysis, though it is incremental as it adapts existing counterfactual explanation techniques to a specific domain.
The paper tackles the problem of interpretability in symbolic music analysis by introducing MUSE-Explainer, a method that generates counterfactual explanations for Graph Neural Network models, showing it provides intuitive insights that can be visualized with standard music tools.
Interpretability is essential for deploying deep learning models in symbolic music analysis, yet most research emphasizes model performance over explanation. To address this, we introduce MUSE-Explainer, a new method that helps reveal how music Graph Neural Network models make decisions by providing clear, human-friendly explanations. Our approach generates counterfactual explanations by making small, meaningful changes to musical score graphs that alter a model's prediction while ensuring the results remain musically coherent. Unlike existing methods, MUSE-Explainer tailors its explanations to the structure of musical data and avoids unrealistic or confusing outputs. We evaluate our method on a music analysis task and show it offers intuitive insights that can be visualized with standard music tools such as Verovio.