Extra colloquium Comuter Science Dr. A. Khodabakhsh, Hasso Plattner Institute (HPI), Potsdam, Germany

Title: Data Science Landscape over Complex Systems

Abstract:

Recent advances in Data Science utilize interdisciplinary approaches and methodologies from Computer Science, Machine Learning, Deep Learning, Statistics, and Engineering for explaining complex systems behavior. The components of these systems are highly connected and their interactions have nonlinear interdependencies. Additionally, the behavior of a complex system demands deeper analysis among spatio-temporal heterogeneous and multi-source data entities. Extracting critical insights from these interactions remains challenging. The motivation of this seminar is to highlight solutions to these scientific problems utilizing Data Science for modeling complex cyber-physical systems and creating novel and effective solutions for empowering decision-making via analytics with an eye on data privacy and explainability. I will talk about 3 use cases of ML data-driven solutions: ML for predictive maintenance in Oil & Gas industry (industrial IoT and sensor data), data privacy in IoT equipped smart homes, and ML for early diagnostics of diseases over medical data.