Submitted Abstract
At the end of 2017, Foyer created the “Data Studio” in order to consolidate its expertise about data exploitation and governance, required by digitalization of client’s experience in insurance business. The objective of “Data Studio” is to develop several projects to exploit this data by Machine Learning and Artificial Intelligence. Our collaboration between Foyer, SnT and the University of Lorraine is one of these projects.Our aim is to take advantage of spatio-temporal dependencies in insurance data to build an innovative mathematical modelling to extract useful information for Foyer (Work Package 1). This would be applied to several use cases, such as predicting customer’s churn or building a recommendation system, expected to be used in production by Foyer (Work Package 2). The implementation of these use cases would be optimized by reducing computation time thanks to GPU or quantum computing, an emerging method which could potentially be a revolution in Machine Learning field (Work Package 3).