Netherlands: PhD Position Probabilistic Tensor Methods

[administrator]

Computer models are an essential tool of modern society. Whether it is for designing airplanes, predicting dominant virus strains in a pandemic or estimating how different policies will impact the CO2 concentration in the next 5 decades, our society makes abundant use of models. While some models can be built from first principles, the majority of artificial intelligence (AI) models are learned from data.

AI models have achieved unprecedented results in the past decade. But this success comes at a cost: it is unsustainable. In fact, the computational power needed to learn large models has doubled every 3.4 months since 2012. In 2019, learning a single model could emit as much carbon as five cars in their lifetimes. This ever-increasing need for computational power is driven by the large amounts of model parameters that can only be reliably learned from both large-scale and high-dimensional data.

The research in this PhD project will be on developing a new theory to make learning models from data sustainable. The key idea of this theory is to significantly compress model parameters with a novel technique: tensor networks. By exploiting correlations tensor networks can capture relevant information such that only a fraction of the original model parameters is required. The focus of this PhD project will be on developing theory for learning kernel machines (support vector machines, Gaussian processes, …) with tensor networks and by using a Bayesian inference approach.

You will join the Delft Tensor AI Lab (DeTAIL) where a team of enthusiastic researchers is developing new tensor theory for applications in machine learning, control and biomedical signal processing.

REQUIREMENTS
Applicants should have:

Completed a relevant MSc degree in an applied sciences field relevant to the PhD research, i.e. Engineering, Computer Science, Statistics, Applied Mathematics;
A strong mathematical background in (numerical) linear algebra, statistics and optimization;
Knowledge of tensor networks/decompositions is a plus;
Good programming skills (Julia, C, Python) is a plus.
Doing a PhD at TU Delft requires English proficiency at a certain level to ensure that the candidate is able to communicate and interact well, participate in English-taught Doctoral Education courses, and write scientific articles and a final thesis. For more details please check the Graduate Schools Admission Requirements.

CONDITIONS OF EMPLOYMENT
Doctoral candidates will be offered a 4-year period of employment in principle, but in the form of 2 employment contracts. An initial 1,5 year contract with an official go/no go progress assessment within 15 months. Followed by an additional contract for the remaining 2,5 years assuming everything goes well and performance requirements are met.

Salary and benefits are in accordance with the Collective Labour Agreement for Dutch Universities, increasing from € 2541 per month in the first year to € 3247 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, discounts 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 Adviceto assist you with your relocation.

https://www.academictransfer.com/en/322815/phd-position-probabilistic-tensor-methods/