Riccardo Guidotti
University of Pisa
From Shapelets and Subsequences to Saliency Maps: Making Time Series Models Understandable in Medicine and Beyond
Abstract Time series are central to many domains, from healthcare monitoring to telemetry data. While obscure machine learning models, especially deep learning, can achieve impressive predictive accuracy, their lack of interpretability hinders adoption in safety-critical settings. This talk explores how shapelets and subsequence-based explanations can bridge the gap between performance and trust. We demonstrate how interpretable subsequences highlight the portions of the signal most responsible for a model’s decision, and how these insights can be extended to also create saliency maps. Through case studies in medicine and traffic safety, we show how subsequence-based interpretable models not only clarifies predictions but also surfaces novel domain knowledge. Finally, by integrating XAI into clinical pipelines, we move toward models that clinicians and practitioners can both trust and act upon.
Riccardo Guidotti is an Associate Professor at University of Pisa. In 2013 and 2010 he graduated cum laude in Computer Science (MS and BS) at University of Pisa. He received the PhD in Computer Science with a thesis on Personal Data Analytics in the same institution. He is currently anAssociate Professor at the Department of Computer Science University of Pisa, Italy, and a member of the Knowledge Discovery and Data Mining Laboratory (KDDLab), a joint research group with the Information Science and Technology Institute of the National Research Council in Pisa. He won the IBM fellowship program and has been an intern in IBM Research Dublin, Ireland in 2015. He also won the DSAA New Generation Data Scientist Award 2018, and the Marco Somalvico Award 2021. His research interests are in explainable artificial intelligence, interpretable machine learning, quantum computing, fairness, and bias detection, time series analysis, data generation, personal data mining, clustering, and analysis of transactional data.