TU Delft

PhD Optimization for and with Machine Learning (OPTIMAL)

Solliciteer op de website van de werkgever

PhD Optimization for and with Machine Learning (OPTIMAL)Faculteit/afdeling: Faculty Electrical Engineering, Mathematics and Computer Science
Niveau: Universitair
Functie-omvang: 36-40 uur per week
Contractduur: 4 years
Salaris: 2325 - 2972 euro per maand (volledige dagtaak) U bekijkt een vacaturetekst oorspronkelijk opgesteld in het Engels.
Delen kunnen beschikbaar zijn in het Nederlands! Faculty Electrical Engineering, Mathematics and Computer Science

The Faculty of Electrical Engineering, Mathematics and Computer Science (EEMCS) brings together three disciplines - electrical engineering, mathematics and computer science. Combined, they reinforce each other and are the driving force behind the technology we use in our daily lives. Technology such as the electricity grid, which our faculty is helping to make future-proof. We are also working on a world in which humans and computers reinforce each other. We are mapping out disease processes using single cell data, and using mathematics to simulate gigantic ash plumes after a volcanic eruption. There is plenty of room here for groundbreaking research. We educate innovative engineers and have excellent labs and facilities that underline our strong international position. In total, more than 1,100 employees and 4,000 students work and study in this innovative environment

Research at Delft Institute of Applied Mathematics (DIAM) is conducted in six research groups: Analysis, Applied Probability, Mathematical Physics, Numerical Analysis, Optimisation, and Statistics. Our research on the construction and analysis of mathematical models related to science and engineering is both fundamental and applied in nature and is often inspired by technical and societal challenges. We have, for instance, tackled complex mathematical problems to develop a model for predicting the flow of ash pollution after a volcano eruption. To come up with innovative solutions we maintain intensive contact with other TU Delft departments, technological institutes and research departments.

Besides research, education is an important cornerstone of our department. We teach the Applied Mathematics BSc and MSc programmes, as well as mathematics courses within other programmes at TU Delft, and national programmes such as “MasterMath”. Together with KTH Royal Institute of Technology (Sweden) and the Technical University of Berlin (Germany) we have also initiated the international joint Master’s programme COSSE (Computer Simulations for Science and Engineering).

We value the great atmosphere at our institute and pride ourselves on our sense of community and the spirit of “getting things done together”, which we combine with mutual appreciation for individual success. To foster interaction and cooperation, we organize lunch seminars where colleagues from the various sections discuss their work, and we arrange an annual social activity where all staff members interact informally to get to know each other in a different setting. Our open and transparent way of communicating is reflected in the way we interact with one another and forms the basis for the way we work together to help each other achieve our goals as a team.

Functie omschrijving

A key component of machine learning is mathematical optimization, that is used, for example, to train neural networks. The goal of this project is to provide new analysis and tools for optimization problems and algorithms arising in machine learning, but also to use insights and tools from machine learning to improve optimization methods. This explains the project title ‘Optimization for and with machine learning’. The project consists of four connected workpackages. The first two workpackages are related to ‘optimization for machine learning’. In the first workpackage we will investigate why the optimization methods currently used in machine learning are often successful in practice and analyze the limits of their computational tractability. The second workpackage is aimed at enhancing the existing optimization algorithms and developing new ones to obtain more accurate machine learning models in an efficient way. The last two workpackages are related to ‘optimization with machine learning’. The third workpackage is aimed at using machine learning to obtain data-centric approximation and optimization algorithms. We will develop algorithms that adapt to the specific data characteristics of the problem instance. The advantage of such data-centric algorithms is more accurate solutions and/or less computation time. In the fourth workpackage we will develop a data-centric optimization modeling approach. In such an approach parts of the resulting optimization model are obtained via machine learning. This data-centric modeling can be used to get more accurate models or can be used in cases where there is no theoretical knowledge available to build the model manually. In addition, we will test our insights on a variety of applications where the consortium members are already involved, including classification problems in the medical sciences, decision problems related to the UN World Food Programme, and routing of shared, self-driving cars.

The institutes and researchers involved in OPTIMAL are

• Tilburg University, Tilburg (Dick den Hertog, Etienne de Klerk)

• CWI, Amsterdam (Nikhil Bansal, Monique Laurent, Leen Stougie)

• Delft University of Technology (Karen Aardal, Leo van Iersel).

Data-centric algorithm design

When we are faced with an optimization problem there are two main alternatives for finding solutions. We can either develop an optimization algorithm for finding a provably best solution, or we can settle for a high-quality solution that is obtained “fast” through an approximation algorithm. For optimization algorithms we will investigate how machine learning can be used to guide the search to an optimal solution. When approximating, it is a challenge to derive algorithms that are not only guaranteed to perform well in the worst-case sense, as is mainly done today, but more interestingly for data that actually occur. We will develop new theoretical concepts for a beyond-worst-case analysis that incorporates data. In addition, we will develop algorithms that have a guaranteed performance according to the developed theory.

Functie eisen

An MSc degree in (Applied) Mathematics, Computer Science or a related discipline, as well as an interest in, knowledge of and some experience with algorithm development, analysis and implementation.

Arbeidsvoorwaarden

TU Delft offers PhD-candidates a 4-year contract, with an official go/no go progress assessment after one year. Salary and benefits are in accordance with the Collective Labour Agreement for Dutch Universities, increasing from € 2325 per month in the first year to € 2972 in the fourth year. As a PhD candidate you will be enrolled in the TU Delft Graduate School. The TU Delft Graduate School provides an inspiring research environment with an excellent team of supervisors, academic staff and a mentor. The Doctoral Education Programme is aimed at developing your transferable, discipline-related and research skills.

The TU Delft offers a customisable compensation package, a discount on health insurance and sport memberships, and a monthly work costs contribution. Flexible work schedules can be arranged.

For international applicants we offer the Coming to Delft Service and Partner Career Advice to assist you with your relocation. An International Children's Centre offers childcare and there is an international primary school.

Informeren en solliciteren

For information about the selection procedure, please contact Drs. Miriam Heemskerk, HR-Advisor, email: vacancies-eemcs@tudelft.nl.

To apply for this PhD positions, the applicant can find more information on the following website: https://leovaniersel.wordpress.com/positions/

The deadline for submission is June 30, 2020.

Acquisitie naar aanleiding van deze vacature wordt niet op prijs gesteld.

Uren per week: 36-40 uur per week

Salaris: € 2.325,- tot € 2.972,- per maand

Solliciteer

Solliciteer op de website van de werkgever

Of solliciteer later
Contactgegevens

Telefoonnummer onbekend
E-mailadres onbekend