Quantifying Explainable AI-introduced signal noise on ECG data with Spectral Entropy
For healthcare AI practitioners, this provides a metric to evaluate the reliability of explainability methods, though it is an incremental application of existing entropy measures.
The paper proposes using spectral entropy to quantify noise introduced by explainability techniques in ECG classification, demonstrating its utility for distinguishing signal from noise in XAI outputs.
Explainability techniques are used to assess the output of various deep learning models. This is especially true in healthcare, where models need to be trusted and decisions justified. Explainability (XAI) tools use heuristics which often add signal noise to the explanation "core". It is not always obvious what is signal from the model and what is noise from the XAI. We propose the use of spectral entropy as a measure of noise in XAI output. We demonstrate its usefulness in the context of classifying arrhythmias in an ECG dataset with different post hoc explainability techniques.