Konstantin Todorov

Associate Professor for AI

Hello!

I am associate professor at the University of Montpellier (maître de conférences HDR, hors classe) and a researcher at the LIRMM laboratory within the ADVANSE team. I received my habilitation in 2020. I used to be a postdoctoral fellow at the Ecole Centrale Paris (France) between 2009 and 2012, after receiving a PhD in Cognitive Science from the University of Osnabrueck (Germany) in 2010. Several years before that, I obtained a MSc (or DEA as they used to call it in France) in Applied Mathematics (option Machine Learning) from the University of Provence in Marseille.

My research activities focus on Artificial Intelligence, with a particular interest in knowledge graphs, NLP, machine learning for graphs and language, knowledge and information extraction. I explore these topics across a variety of application domains, including journalism, biology, immunogenetics, and cultural heritage. Some of the keywords that charactirize my research are interdisplinarity (via collaborations with the social sciences, media & cultural institutions), societal cohesion (e.g. via fighting disinformation), public health (e.g. via supporting vaccine research), open and inclusive science.

At LIRMM, I’m co-leading the AI and Data Science transversal axis. I’m also the lab’s European correspondent.

News

Teaching, Research & Supervision

Teaching

I am co-head of the ICo (Intégration de Compétences) masters programme at the Faculty of Science at the University of Montpellier. I’m mainly teaching at masters 1 and 2 level courses on machine learning, statistics and probabilities and semantic technologies.

Publications

I publish and review for venues such as The Web Conference (WWW), the International Semantic Web Conference (ISWC), the Extended Semantic Web Conference (ESWC), the International Conference on Information and Knowledge Management (CIKM), ACM Hypertext (HT), the Semantic Web Journal (SWJ), the ACM Transactions of the Web, Briefings of Bioinformatics. Most of my publications are indexed by Google Scholar (H 17) and DBLP.

Funded projects

I am currently coordinating the French-German ANR project AI4Sci and the ANR PRC project DACE-DL.  AI4Sci detects and analyses science-related discourse online (namley on X/Twitter), with applications related to fake-news and disinformation detection. DACE-DL is building a foundational model for entity alignment across knowledge graphs by leveraging a neuro-symbolic AI approach, combining Natural Language Inference with graph neural models. I’m also co-leading the FEDER IA-EMOTIONS project, which analyses emotions dynamics in conversations with generative AI chatbots. I’m a member of the ANR DIG-AI project. In the past, I used to co-lead the ANR DOREMUS project and was part of several ANR projects, such as D2KAB or PratikPharma.

Knowledge graphs

Some examples of resources that I have developped in collaboration with colleagues from different fields include: ClaimsKG, the first knowledge graph of fact-checked claims; AgroLD – a knowledge graph for the plant sciences; IMGT-KG – a knowledge graph for immunogenetics, DOREMUS – a graph of linked music works, etc.

Partners

Some of my recent collaborators include: GESIS (Germany), FORTH (Greece), University of Trier (Germany), LERASS (Montpellier), IRIT (Toulouse), MISTEA (Montpellier), IRD (Montpellier), IGH (Montpellier), FZD (Germany), Ex.AI (Spain), BnF (Paris), Radio France (Paris), SciencePo MediaLab (Paris).

PhD Students

  1. Mohamed Ben Ellefi
    Defense year: 2016; Thesis title: Profile-based Dataset Recommendation for Data Linking

  2. Manel Achichi
    Defense year: 2018; Thesis title: Linking and Publishing Heterogeneous Open Data in the Musical Field

  3. Abdelnasser Tigrine
    Abandoned in 4th year (no defense). Thesis title: Using Background Knowledge for Facilitating the Ontology Matching Task

  4. Anjana Khushwaha
    Defense year: 2023; Thesis title: Prediction and analysis of epitope and paratope interaction

  5. Gaoussou Sanou
    Defense year: 2024; Thesis title: Une base de connaissance en immunogénétique pour la découverte de la nouvelle connaissance scientifique

  6. Baptiste Darnala
    Defense year: 2024; Thesis title: Hybridation des méthodes d’apprentissage et Web sémantique pour l’optimisation et la planification de cultures maraîchères en agro-écologie 

  7. Salim Hafid
    Defense year: 2024; Thesis title: Detection, linking and interpretation of science-related claims
    and their contexts from online discourse 

  8. Ensiyeh Raoufi
    Planned defense: 2025; Thesis title: AI-driven Bottom-Up Data Linking: from Knowledge Graph Profiling to Meaningful and Interpretable Links

  9. Florian Lecourt
    PhD started: 2024 (ongoing); Thesis title: Emotion detection in dialogues between humans and AI chatbots

Contact

todorov [at] lirmm.fr