Thèse Neuro-Symbolic Data Integration For Multi-Modal Biomedical Data H/F - Doctorat.Gouv.Fr
- CDD
- Doctorat.Gouv.Fr
Les missions du poste
Établissement : Nantes Université École doctorale : École doctorale Mathématiques et Sciences et Technologies du numérique, de l'Information et de la Communication Laboratoire de recherche : Laboratoire des Sciences du Numérique de Nantes Direction de la thèse : Pascal MOLLI ORCID 0000-0001-8048-273X Date limite de candidature : 2026-11-30T00:00:00
In this PhD proposal, we aim to explore the potential of Language Models (LMs) for multi-modal biomedical data integration, including scientifi c papers and tabular experimental data. We will develop neuro-symbolic data integration methods for generating data schemas, formulating queries, and ensuring the trustworthiness of results through explanations for the generated answers.
Computer, office, traveling for conferences
The PhD will address the following key research questions:
How can we combine symbolic querying (logical/graph-based) with neural (LMs) retrieval methods for scalable and privacy-preserving question answering?
How can KGs act as semantic indexes over large experimental datasets, enabling efficient retrieval of relevant results?
How can neuro-symbolic approaches ensure that generated answers are verifiable and interpretable?