Achille Thin

Senior Lecturer in Statistics and Artificial Intelligence, I apply my research to the agricultural and agronomical sectors and teach students from AgroParisTech tools for statistical modelling.

Developing novel efficient algorithms and methods for various tasks in Bayesian Inference and Deep Learning: Variational Auto Encoders building on Normalizing Flows, Langevin dynamics or Markov chain Monte Carlo; Non reversible Markov chain Monte Carlo framework for efficient high dimensional sampling; New Markov chain Monte Carlo algorithm and novel approaches to Normalizing flows; ... Application to various tasks, including for example image generation, recommender systems, uncertainty quantification, and unsupervised learning in general. 2nd Place on NeurIPS 2021 international competition for Approximate Inference in Bayesian Deep Learning. PhD defended the 10th of June 2022 ! Publications in at NeurIPS and ICML conferences. PhD directors: Arnaud Doucet (University of Oxford) and Éric Moulines (Ecole polytechnique) Collaboration with Maxim Panov and Nikita Kotelevskii (Skoltech), Sylvain Le Corff and Yazid Janati (Telecom Sud-Paris), Christophe Andrieu (University of Bristol), Christian Robert (Université Paris-Dauphine, University of Warwick).

Lecturer for the Master of Science "Big Data for Business" at Ecole polytechnique and HEC Paris. Teaching a class of ~30 Master students mathematical and statistical tools for reasoning, decision taking and analysis.