The OPTIMAL project is run by a consortium of researchers at University of Amsterdam (Dick den Hertogi, coordinator), CWI, Amsterdam (Monique Laurent, Guido Schaefer, Leen Stougie), TU Delft (Karen Aardal, Leo van Iersel), and Tilburg University (Etienne de Klerk). The consortium has 5 PhD positions and 6 postdoc positions in total, including one PhD position and 2 postdoc positions at CWI.

There is one postdoc position available at CWI with Monique Laurent.

Research topic

Optimization plays a central role in machine learning. The objective of this research is enhancing existing optimization algorithms and developng new ones to obtain more accurate machine learning models. The overall goal is to exploit structural properties of data, which can be geometric, algebraic or combinatorial, for the design of dedicated solution approaches. Special areas of focus include (but are not limited to):
- Investigating the use of polynomial functions in machine learning. In particular, how to use tools like the sums-of-squares method and the general framework of the moment problem for the design of efficient approximations for machine learning tasks involving polynomials and geometric problems such as detecting global shapes of point clouds or outliers through properties of their unknown distributions.
- Investigating combinatorial algorithms for detecting structural properties of data points, such as an unknown linear ordering, when the points are given through their pairwise similarities.

Requirements

We are looking for a candidate with a strong mathematical background and affinity with algebraic, combinatorial and geometric methods and with expertise in optimization (semidefinite prgramming), background knowledge in machine learning is a big advantage. A PhD degree in mathematics (or operations rsearch, with a strong mathematical background) is required.

For information about this position please contact Monique Laurent (monique at cwi.nl).

Research environment

The researcher will be embedded in the research group Networks and Optimization at CWI. He/she will benefit from a rich range of scientific acivities, within CWI and the OPTIMAL consortium, as well as possible collaboration within the two Marie-Curie ITN consortia MINOA and POEMA. For an overview of CWI research see the research groups and the general CWI webpage.

Terms and conditions

The position is for one year, with possible one year-extension, starting from fall 2021 or later. The position will remain open until a suitable candidate will be found.

More details about the position and on how to apply can be found here (site will open soon). You may also send a CV, motivation letter, link to your papers and PhD dissertation, and a list of three references to monique at cwi.nl