• Explore the Mathematics behind Machine Learning

Research

Explore the Mathematics behind Machine Learning

L2F’s scientific ambition is to discover the fundamental laws of Machine Learning by developing its underlying mathematical theory. We apply Topology and Geometry to shed a new light on Data Science, Natural Language Processing, and Artificial Intelligence. With PhDs in Mathematics, Theoretical Physics, and Computer Science, L2F's Research Department provides an exciting environment for multi-disciplinary researchers. Our presence on the EPFL campus enables us to exchange ideas with world-leading experts.

L2F’s research project “Topological Warning Signals for Critical System Transitions” has recently been accepted by Innosuisse (Swiss Innovation Agency), in partnership with Prof. Kathryn Hess from the Laboratory for Topology and Neuroscience (EPFL) and with Prof. Alberto Dassatti from HEIG-VD. Inspired by critical phase transitions, our first goal is to define a new type of early warning signals able to better anticipate sudden changes of regime such as machine failures, financial crashes, earthquakes, or epileptic seizures. The second goal concerns the creation of new local features from the topology of datasets, in order to generalize current Topological Data Analysis techniques and enhance the performance of Machine Learning models.