Professor Olivier Sylvain (Fordham Law School): “Recovering Tech's Humanity”

Laurent is a senior CNRS Research Engineer in Data Science. He currently works half-time at the Mathematics Institute of Toulouse, and half-time at the 3IA Artificial and Natural Intelligence Toulouse Institute. He has more than 15 years of experience in Image Analysis, mainly in medical imaging and has developed these 7 last years a strong experience in Machine Learning with broad applications. He has published over 70 communications in international scientific journals and peer-reviewed proceedings, as well as two book chapters. During his career, he obtained grants from e.g. ANR, INSERM, ERC, Region Midi-Pyrénées, FRM. He developed or supervised the development of 5 packages in Python, C++ and Java which are freely available on sourceforge and github. He also supervised 30 students and engineers in applied mathematics and computer science from various universities and engineer schools such as Univ. Toulouse, INSA Toulouse, INP ENSEEIHT, ENS Lyon, ENS Cachan, Imperial College London, IUT Toulouse (L3, M1, M2, Phd levels and postdocs/engineers). He has taught the fields of Machine Learning, Image Analysis, Statistics, Optimization, GPU computing, Computational Fluid Mechanics at Univ. Toulouse, ISAE-Supaero , INSA Toulouse, Toulouse School of Economics, Mohammed VI Polytechnic University (through Ecole Polytechnique) and Ho-Chi-Minh city University of Science. He finally obtained in 2019 a 3IA ANITI Chair « Law, Accountability and Social Trust in Artificial Intelligence » to establish novel links between Artificial Intelligence and Law, in collaboration with C. Castets-Renard (Univ. Ottawa - Law) and S. Cussat-Blanc (Univ Toulouse - Computer Science). Previously, he worked in three research institutes after having defended his PhD thesis in engineering science (specifically 3D image analysis applied to fluid mechanics). He first worked at CEA Saclay/Neurospin, where he developed novel regularization strategies for the analysis of 3D+t fMRI time series. At Imperial College London and University of Oxford, he then developed application-motivated strategies to make it possible to efficiently take advantage of properties of the mathematically-grounded LDDMM image registration formalism on 3D medical images.