Projects

Current projets

  • H2020 STARWARS (2022 - 2026) : European project starting soon

  • ANR PRC CROQUIS (2022 - 2025) : Collecting, Representing, cOmpleting, merging and Querying heterogeneous and UncertaIn waStewater and stormwater network data

  • Prédiction et optimisation de ressources aéroportuaires (2021- 2023, contrat industriel ADP and Exakis Nelite)

I've been involved and in charge of the following projects :

H2020-MSCA-RISE-2015 AniAge (Jan. 2016 - Dec 2019)

AniAge: High Dimensional Heterogeneous Data based Animation Techniques for Southeast Asian Intangible Cultural Heritage Digital Content. Link

AniAge is a multidisciplinary project involving researchers in Artificial Intelligence and in Computer Animation. The project covers three main research topics: i) developing new digital animation techniques, ii) management of large and heterogeneous data and iii) an application to the intangible cultural heritage (IHC) of Southeast Asia countries. Contributions from CRI concern mainly the management of large and heterogeneous data. The consortium consists of six partners: NCCA (National Centre for Computer Animation, Bournemouth University, UK) who is the project coordinator, CRIL CNRS UMR 8188 et Université d’Artois, HMI (Human Machine Interaction Lab at Vietnam National University, Vietnam), CAMT (College of Arts Media and Technology at Chiang Mai University, Thailand), CICT (College of ICT at Can Tho University, Vietnam) and ViCube (Vision, Virtual Visualization Lab at and Universiti Teknologi Malaysia).

CRIL contributed in three main tasks :

  • Data enrichment through video annotation :

      • Sylvain Lagrue, Nathalie Chetcuti-Sperandio, Fabien Delorme, Ma Thi Chau, Duyen Ngo Thi, Karim Tabia, Salem Benferhat: An Ontology Web Application-based Annotation Tool for Intangible Culture Heritage Dance Videos. SUMAC @ ACM Multimedia 2019: 75-81

      • Truong-Thanh Ma, Salem Benferhat, Zied Bouraoui, Karim Tabia, Thanh-Nghi Do, Nguyen-Khang Pham: An Automatic Extraction Tool for Ethnic Vietnamese Thai Dances Concepts. ICMLA 2019: 1527-1530

video3.mp4

  • Querying large-scale datasets and Knowledge modeling using ontologies :

      • Sihem Belabbes, Yacine Izza, Nizar Mhadhbi, Tri-Thuc Vo, Karim Tabia, Salem Benferhat, An Ontology-based Approach for Building and Querying ICH Video Datasets, In 12th International Conference on Agents and Artificial Intelligence ICAART, 2020.

      • Sihem Belabbes, Chi Wee Tan, Tri-Thuc Vo, Yacine Izza, Karim Tabia, Sylvain Lagrue, Salem Benferhat, Query Answering from Traditional Dance Videos: Case Study of Zapin Dances, In 31st IEEE International Conference on Tools with Artificial Intelligence, ICTAI 2019, 1638-1642, Portland, OR, USA, November 4-6, 2019.


Aviesan OCIP-Nut: Objets Connectes Intelligents et Personnalises dans le domaine de la Nutrition. Link

This project involved CRIL and LIMICS (INSERM, UMRS 1142) and it aimed to carry out preliminary studies and research in order to identify scientific obstacles and target feasible and high-impact avenues concerning intelligent connected objects in the field of nutrition. To do this, it is necessary to explore several issues. Here are the most relevant:

  1. How to use data from different connected sensors and objects?

  2. How to exploit and take into account user constraints and preferences in recommendation applications on devices such as connected objects?

  3. How to use the history data of a user to prevent certain problems such as nutritional deficiencies?

Outcomes of this project were published mainly in the following papers :

  • Cécile Carra, Karim Tabia, Smart Home for Seniors: Opportunities and Challenges for AI, In 12th International Conference on Agents and Artificial Intelligence ICAART, 2020.

  • Karim Tabia, Hugues Wattez, Nicolas Ydée, Karima Sedki: Data Analytics and Visualization for Connected Objects: A Case Study for Sleep and Physical Activity Trackers. IEA/AIE 2018: 685-696

  • Thibaut Vallee, Karima Sedki, Sylvie Despres, M.-Christine Jaulant, Karim Tabia, Adrien Ugon: On Personalization in IoT. CSCI 2016, Volume: 1, Pages: 186-191, 2016, DOI:10.1109/CSCI.2016.0042

PEPS FasciDo MAPPOS : (2015)

Inference in interval-based Bayesian networks with an application in pattern recognition. MAP queries are of particular importance since they are the basis of several applications such as classification, diagnostics, etc. This problem is all the more important since there are no efficient algorithms to respond to these requests. The proposed approach aims to represent the family of compatible Bayesian networks by a single possibilistic network and respond to queries using the latter.

This project is carried out in collaboration with the DuKe team / LS2N CNRS UMR 6241.

  • Maroua Haddad, Philippe Leray, Amélie Levray, Karim Tabia: Learning the Parameters of Possibilistic Networks from Data: Empirical Comparison. FLAIRS Conference 2017: 736-741

  • Salem Benferhat, Amélie Levray, Karim Tabia: Approximating MAP Inference in Credal Networks Using Probability-Possibility Transformations. ICTAI 2017: 1057-1064

  • Karim Tabia: Possibilistic Graphical Models for Uncertainty Modeling. SUM 2016: 33-48

  • Salem Benferhat, Amélie Levray, Karim Tabia: On the Analysis of Probability-Possibility Transformations: Changing Operations and Graphical Models. ECSQARU 2015: 279-289