PhD position at IRIT (Toulouse) in Opinion analysis and synthesis

PhD student position at IRIT (Toulouse) and CEA:

Opinion analysis and synthesis:  Exploiting spatio-temporal and social contexts

Description:

The large amount of information generated in social networks brings new opportunities and new applications, interesting for some of them in terms of markets, and very useful and important for others in terms of security (cyber security).

The topic of this thesis deals with cyber security and it mainly concerns opinion analysis. Our goal is to develop solutions (approaches) to automatically detect and synthesize opinions about a given subject. We will be interested on opinions (comments, feedbacks) about people or event by considering their spatiotemporal context.

Opinion analysis is an active area where numerous approaches, including those based on NLP (Natural language Processing), opinion lexicons or machine learning have been proposed to detect the polarity (subjectivity of text and its polarity (positive, negative, neutral)) of a text. Beyond simply assigning a polarity opinion of a text, a more detailed analysis is often required to have a better characterization of the opinion, for instance find out which object or which facets/aspects of the object are concerned by the opinion. Although several approaches have been proposed including rule-based models or statistics, some research issues remain open, especially when we deal with short texts such as tweets and when we consider the social and the spatio-temporal contexts of the concerned topic. These later, particularly, the spatio-temporal context, is of great importance when we aim at building a summary or a synthesis of opinions about a topic.

The aim of this thesis is to develop and evaluate approaches that are capable to produce a consolidated opinion summary/synthesis about a given topic (event, person) by considering the social and the spatio-temporal contexts surrounding the concerned topic (event).

Candidate Profile:  Applicants should have an excellent degree (M.Sc., M.Eng. or equivalent) in Computer Science.
Application: Detailed CV, academic records, motivation letter and recommendation letters , to be sent to Mohand Boughanem (boughanem@irit.fr)
Supervision : The PhD student will be enrolled at the University Paul Sabatier in Toulouse and will conduct his research in IRIT lab. in Toulouse (Université Paul Sabatier-IRIT, 118 Route de Narbonne 31062).
Supervisors:  Mohand BOUGHANEM – IRIT and  Romaric BESANCON – CEA Tech LIST/DIASI/LVIC
Funding: Doctoral allowance funded by "Contrat Laboratoire Entreprise CEA-IRIT

the  Monthly salary, approximate 1,600 Euros, Net.

Duration : 36 mois

Starting Date: 1 October (or 1 November) 2015