PhD position: fuzzy rough set models for machine learning and sentiment analysis

Applications are invited for a fully-funded 4-year PhD scholarship within the Computational Web Intelligence (CWI) lab. CWI is a unit of the Fuzziness and Uncertainty Modelling (FUM) research group of the Department of Applied Mathematics, Computer Science and Statistics (TWIST) at Ghent University.

 

The scholarship forms part of the project “Fuzzy sets and rough sets for machine learning, sentiment analysis and web intelligence” funded by the Odysseus programme of the Research Foundation-Flanders. Fuzzy sets and rough sets are well-established mathematical AI models designed to deal with imperfections in data: on the one hand, fuzzy sets model vague information, by recognizing that membership to certain concepts, or logical truth of certain propositions, is a matter of degree. On the other hand, rough sets deal with incomplete information by approximating concepts from below and from above, using information granules as building blocks. Here, we focus on the approximation of concepts using a fuzzy relation expressing degrees of similarity or preference.

 

In particular, the PhD scholarship involves harnessing and combining a variety of machine learning (ML) techniques (deep learning, similarity learning, etc.) and data settings (big data, semi-supervised, multi-instance, multi-label, etc.) with improved or novel fuzzy rough set models. It will build on the results of previous research conducted at CWI, and at the same time, will reach out to (aspect-based) sentiment analysis as a test bed for the developed methods.

 

Your key responsibilities include:

 

• Designing new ML algorithms based on fuzzy rough sets;
• Setting up experimental studies with benchmark data to evaluate your algorithms;
• Working together with other PhD students and senior researchers involved in the project;
• Reporting and presenting your research in top-tier international journals and conferences;
• Successfully defending a doctoral dissertation at Ghent University;
• Contributing (to a limited extent) to teaching and community engagement activities within the department.

 

We are especially interested in candidates with the following profile:

 

• A Masters’ degree in Computer Science or Mathematics;
• Very good to excellent study grades, as evident from university transcripts;
• Excellent programming skills (e.g. Java, Python, …);
• Experience with Big Data frameworks (e.g. Hadoop, Spark) and Deep Learning libraries (e.g. TensorFlow, PyTorch) is a plus;
• Good knowledge of machine learning, artificial intelligence and mathematical modeling;
• Strong motivation and aptitude to tackle challenging research problems and to complete the PhD within the prescribed period;
• Good interpersonal and team-playing skills;
• Proficiency in English and excellent writing skills;
• Experience with natural language processing and/or soft computing is a plus.

 

We offer:

 

• A four-year full-time (100%) PhD student position;
• A competitive salary as set out by the Flemish Government;
• A dynamic, multi-disciplinary and ambitious research team with a wide international network (http://www.cwi.ugent.be);
• Expertise, state-of-the-art research infrastructure;
• A four-year Doctoral Training Program at Ghent University

 

 

How to apply

 

The scholarship’s prospective starting date is October 1, 2018. Interested candidates should e-mail:

 

1. a detailed curriculum vitae;
2. a motivation letter;
3. BSc and MSc transcripts of grades and the MSc thesis;
4. Two reference letters

 

to Dr. Chris Cornelis (Chris.Cornelis@UGent.be) and Dr. Daniel Peralta (Daniel.Peralta@UGent.be) by May 31st, 2018. Indicate “Application: PhD Researcher on fuzzy rough set models for machine learning and sentiment analysis” in the email subject.

 

Note that all applications will be thoroughly screened in order to satisfy the high scientific standards of our research group. The selection process will take place along June 2018. As part of this process, potential candidates will have to conduct a small research task to assess their programming and problem solving skills, and will be invited for an interview (possible by Skype).