Sort by
Refine Your Search
-
minimizing error and maximizing efficiency, is computationally challenging—no known polynomial-time algorithm exists to solve it optimally in all cases. Because of this complexity, researchers typically rely
-
techniques and the structure of bilevel problems in large-scale settings. Objectives The goal of this postdoctoral project is to develop scalable blackbox optimization algorithms tailored to bilevel problems
-
on stochastic Riemannian optimization algorithms, these methods still suffer from limitations in computational complexity. The post-doctoral fellow will build upon this preliminary work to investigate
-
algorithms for dynamic structured data, with a particular focus on time sequences of graphs, graph signals, and time sequences on groups and manifolds. Special emphasis will be placed on non-parametric
-
suitable data models [CSC+23]. Objectives As far as the design of efficient numerical algorithms in an off-the-grid setting is concerned, the problem is challenging, since the optimization is defined in
-
in developing new tools to understand the nervous system and to explore theories behind neural phenomena. As for developing new tools, we have been working on network alignment algorithms [FCC+21] and
-
train robust machine learning (ML) algorithms without exchanging the actual data. The benefits of such a decentralized technology over personal and confidential data are multiple and already include some
-
revisit discretization methodologies in view of modern requirements and computational capabilities. The candidate will focus on developing mesh generation algorithms meeting the following criteria
-
. The monitoring of telecommunications and energy production and distribution networks are characteristic examples of such time-critical applications. The project aims to propose unsupervised online CPD algorithms
-
, sensor failures, or the aggregation of datasets from multiple sources. There is a rich literature on how to impute missing values, for example, considering the EM algorithm [Dempster et al., 1977], low