ETRO VUB
About ETRO  |  News  |  Events  |  Vacancies  |  Contact  
Home Research Education Industry Publications About ETRO

Master theses

Current and past ideas and concepts for Master Theses.

Time-aware Matrix and Tensor completion

Subject

Matrix Completion concerns the problem of recovering a full matrix given a small portion of observed entries. While Matrix Completion concerns only matrices (or second-order tensors), Tensor Completion generalizes the problem to deal with tensors of any order. Both Matrix Completion and Tensor Completion have various applications across Machine Learning and Signal Processing domains. One notable example is Recommender System, which is one of the most important engines behind big companies like Amazon, eBay and Netflix. Existing works in this topic focus mainly on static recommendations, utilizing fixed matrix factor models. Recently, some efforts have been spent on modelling the temporal dynamics in recommender systems, yet, their results still need to be improved. This thesis will focus on time-aware recommender systems, utilizing novel matrix and tensor completion models to capture the recommendation variations across the temporal domain. The thesis will be based on the core deep neural network models for inverse problem, which have been developed at VUB-ETRO.

Kind of work

The thesis will consist of the following steps:
(1) Literature study
(2) Re-implementing state-of-the-art algorithms, with the focus on deep-neural-network-based algorithms
(3) Evaluating and analyzing strong points and drawbacks of state-of-the-art algorithms on various standard datasets
(4) Improving over the state-of-the-art algorithms
(5) Writing the thesis

Framework of the Thesis

Matrix Completion and more generally Inverse Problem have been one of the research focuses at VUB-ETRO. The work within this thesis will go fully in line with research in these themes at VUB-ETRO. Publication of the research results in international conferences and journals is expected and fully supported.

Number of Students

1

Expected Student Profile

- Computer Science / Electrical Engineering / Mathematics
- Good programming skills in Python
- Good knowledge of machine learning

Promotor

Prof. Dr. Ir. Nikolaos Deligiannis

+32 (0)2 629 1683

ndeligia@etrovub.be

more info

Supervisor

Mr. Duc Nguyen

+32 (0)2 629 1686

mdnguyen@etrovub.be

more info

- Contact person

- IRIS

- AVSP

- LAMI

- Contact person

- Thesis proposals

- ETRO Courses

- Contact person

- Spin-offs

- Know How

- Journals

- Conferences

- Books

- Vacancies

- News

- Events

- Press

Contact

ETRO Department

info@etro.vub.ac.be

Tel: +32 2 629 29 30

©2019 • Vrije Universiteit Brussel • ETRO Dept. • Pleinlaan 2 • 1050 Brussels • Tel: +32 2 629 2930 (secretariat) • Fax: +32 2 629 2883 • WebmasterDisclaimer