Graduation Project: Estimation of Information-theoretic Quantities via GPU

Your challenge

We are looking for an enthusiastic candidate who is passionate to perform research in communication and information theory applied to fiber optical and wireless communications.

Project description

Within our team, we are looking at methods to efficiently compute information-theoretic quantities (such as mutual information and generalized mutual information) in high-dimensional spaces. Such problems appear in telecommunications when dense constellations are used in high-dimensional spaces (many antennas, modes, cores, etc.). The main objective of this project is to study the applicability of commercial state-of-the art (NVIDIA) GPUs usually used for gaming (e.g., the NVIDIA Titan X) as a tool to efficiently calculate information-theoretic quantities. These quantities will later be used to characterize telecommunication systems.

Your team

You will be working in the Signal Processing Systems (SPS) Group, TU/e Eindhoven. This project is also in closed collaboration with the Electro-optical Communications (ECO) Group, where a multi-terabit fiber optical test-bed is available for experiments. You will interact with experts in communication theory, information theory, and fiber optics.

Your responsibilities

Together with a team in the SPS and ECO Groups, you will:

  • Increase your theoretical and practical understanding of some information-theoretic quantities,
  • Develop numerical routines to efficiently numerically calculate such quantities, and
  • Evaluate their performance via simulations and (fiber optical) experiments.


We are looking for

To be successful in this project we are looking for a student with the following profile:

  • Working towards a MSc. in Electrical Engineering, Telecommunication Technology, Computer Science, Embedded Systems, Applied Mathematics, or equivalent.
  • Good knowledge in communication and/or information theory.
  • Interest in wireless and/or optical communications as well on CUDA programming.
  • Pro-active and self-motivated with at least 6 months’ availability.
  • Language: Good written and spoken command of English is mandatory.
  • Previous internship experience not necessary.


Notes

Required documents:

  • Student registration form (Proof of enrollment at the current education)
  • CV
  • Short cover letter (max 1 page) outlining your motivation and informing your availability 


For interested applicants, please contact Dr. Alex Alvarado (a.alvarado@tue.nl)