Sara Scaramuccia

Research Affiliate - Topological data analysis, Computational topology, Machine learning, Topological visual analytics

Website: link

My current project is mainly rooted in topological data analysis thus aiming at developing new methods in computational topology to provide qualitative understanding of data. The project is developed in partnership with the Department of Mathematical Sciences of the Politecnico di Torino.

My role is that of exploiting algebraic topology tools in combination with learning theory, exploring the mathematical foundations in the interplaying between Machine Learning and Artificial Intelligence.
My research finds applications in several industrial problems, e.g. anomaly detection in mechanical and civil engineering, smart building modeling and simulation, analysis of systems of fractures coming from geophysics.

From 2019 to 2020, I have been a postdoc fellow at the Université de la Côte d’Azur. My project was in parternship with Amadeus SA and it consisted in developing new clustering algorithms guided by user feedbacks, to optimize clustering quality in the specific domain of flight recommendations by means of Active Learning techniques.

In 2018, I got  my PhD  in Computer Science at University of Genova, Italy.
The project was focused on computational methods and data structures for Topological Data Analysis:
optimized algorithms for the retrieval of multidimensional persistent homology invariants
multidimensional segmentation of domains via discrete Morse theory.