An Empirical Study of Explainable AI Techniques on Deep Learning Models For Time Series Tasks

Decision explanations of machine learning black-box models are often\ngenerated by applying Explainable AI (XAI) techniques. However, many proposed\nXAI methods produce unverified outputs. Evaluation and verification are usually\nachieved with a visual interpretation by humans on individual images or text.\nIn this preregistration, we propose an empirical study and benchmark framework\nto apply attribution methods for neural networks developed for images and text\ndata on time series. We present a methodology to automatically evaluate and\nrank attribution techniques on time series using perturbation methods to\nidentify reliable approaches.\n