High-Dimensional Inference via Convex Optimization
DESCRIPTION
Our focus is on new techniques—theoretically and computationally sound—for a broad class of statistical problems, using a new “operational” framework for analyzing inferential routines that relies on the extensive use of the methodology involved in developing modern optimization.
Our research agenda includes, but is not limited to,
Design of Efficient Testing Routines with Applications to Diagnostics Based on Heterogeneous Data
New methods for recovering signals from nonlinear observations, with applications in classification, identification, and inverse problems for nonlinear PDEs
stochastic optimization using indirect observations (e.g., privacy-protected data)
design of robust procedures with a focus on iterative (adversarial) adaptive techniques
large-scale online algorithmic implementation of inference routines
adaptive techniques for network inference from multisensor data, application to the identification of temporal dynamics in biomedical data, (stochastic) optimization methods that enable the efficient use of streaming data, …
ACTIVITIES
The chair is collaborating with the French companies BioMérieux and ST Microelectronics through the supervision of two CIFRE PhD students. We are continuing our established collaborations with the “MAGNET” chair (through E. Devijver’s joint membership) and the chairs “Towards More Data Efficiency in Machine Learning” and “Optimization & Learning.” New collaborations have been established with ENSAE-CREST through the joint supervision of PhD projects.
- Efficient statistical routines based on convex optimization
- Algorithms for stochastic and deterministic optimization
- Stochastic system identification and time series analysis
- Causal and functional data inference, semi-supervised and unsupervised learning
- Interdisciplinary collaborations, specifically in healthcare and electronic circuit design
Share the linkCopyCopiedClose the modal windowShare the URL of this pageI recommend this page:Available at this address:The page will then be accessible from your "My Favorites" menu.Stop the videoPlay the videoMutePlay audioChat: Have a question?Robo Fabrica ChatbotMatomo Traffic StatisticsX (formerly Twitter)
