The Institut de Recherche en Informatique de Toulouse (IRIT-Toulouse University) has a 9 months position offer within the INTACT project on using NLP for the detection of intention to acts from social media during crises. The project brings together academic from linguistics and computational linguistics as well as expert in crisis management. The offer is detailed below.

**Job description**

Social media platforms offer great opportunities to identify public  commitments to intentions and de-sires/plans to act from user-generated contents. Automatically  detecting intentions can have many potential applications, such as in commerce marketing (consumption intentions: sell, purchase), securi-ty and defense (intention to attack, to connect to terrorist organizations), health (intention to suicide), and  emergency management (intentions to help, evacuate). Mining intentions from texts can thus help decision makers to differentiate between intentional and non-intentional messages ("I want to buy this great phone" vs."Trump won the US election") and to better identify the type of intention behind each message which is a primordial step to help companies, government, or institutions to better predict users' future actions and thus improve their strategies.

The position is funded by the French Interior Department (no French citizenship required) within INTACT, a joint project between Institut de Recherche en Informatique de Toulouse (IRIT-Toulouse University) and Institut Jean Nicod (IJN-Ecole Normale Supérieure).

The candidate will investigate intention detection in social media texts (tweets) during crisis events (e.g., hurricane, earthquake, flooding, etc.). The work will build on (Kozlowski  et al, 2020) and will focus in particular on: (1) Developing deep learning models for an accurate extraction of public commitments to act, (2) Exploring domain adaptation techniques to extend the models to other types of crises (e.g., sudden events like building collapse, sanitary crises).

Diego Kozlowski, Elisa Lannelongue, Frédéric Saudemont, Farah Benamara, Alda Mari, Véronique Moriceau, Abdelmoumene Boumadane. (2020). A three-level classification of French tweets in ecological crises. Inf. Process. Manag. 57(5).

**Additional information**

-->Duration of the position: 9 months
--> Salary : 2400 euros (gross salary)
--> Starting date: September 2021 (negotiable for an early start)

**Ideal Candidate**

-->PhD in Computational Linguistics or Master in Computer science
--> A very good experience in deep learning approaches for NLP
--> Good programming skills in Python
--> Good knowledge in linguistic description for NLP.

**Application**
Send your CV together with a list of publications and a statement of research interests to Farah Benamara (farah.benamara@irit.fr) and Véronique Moriceau (veronique.moriceau@irit.fr) before April 30th 2021