# Matinée des (post-)doctorant.e.s de l'axe Santé Numérique du Labri. **L’édition 2024** se déroulera le **mercredi 20 mars 2024** dans l'**amphi du LaBRI** de 9h30 à 12h15 et se clôturera par un buffet. ## 9h30 : Séminaire d'ouverture de la matinée - **Victoria Bourgeais**, Team: BKB, **Enseignante-Chercheuse de l'université de Bordeaux.** **Titre :** "Explainable AI for Personalized Medicine" **Résumé :** "Deep learning has been a significant breakthrough in artificial intelligence, particularly excelling in tasks like image analysis and natural language processing. However, applying this technology to healthcare presents challenges due to its lack of interpretability. Often seen as a "black box," deep learning models offer little explanation for their predictions. Yet, interpretability is essential for practical use, especially in sensitive fields like medicine. It ensures that the model's predictions are based on reliable patient representations, not irrelevant artifacts in the data. Moreover, interpretation could lead to new biological discoveries by identifying relevant data signatures. In my talk, I will introduce our approaches for interpreting neural networks used to aid diagnosis based on patient gene expression data or electrocardiograms. Interpretation is achieved by integrating domain knowledge as concepts, either a posteriori or a priori." ## 10h15 : Présentations Flash des (post-)doctorant.e.s ***Département Systèmes et Données (SeD)*** - **Yanis Asloudj**, Team BKB et AHeaD, Inserm U1219 **Titre :** "Single-cell RNA-seq analysis under the frequent itemset framework". - **Marina Boudin**, Team AHeaD, Inserm U1219 **Titre :** "The OREGANO knowledge graph for computational drug repurposing". - **Elsa Claude**, Team BKB et ADLab (Canada) **Titre :** "How to use hybrid ensemble feature selection to robustly identify disease biomarkers: example of Stage IV colorectal cancer with RNAseq data" - **Antonin Colajanni**, Team SISTM et BKB **Titre :** "Detection of non-human sequences in human RNA-sequencing" ## 10h 45 : Pause café ## 11h15 : Présentations Flash des (post-)doctorant.e.s ***Département Méthodes et Modèles Formels (M2F)*** - **Maeva Andriantsoamberomanga**, Team: Neuromodulation and Neuroprosthetics, INRIA **Titre :** "Modélisation computationnelle de l'hippocampe pour étudier les paramètres de stimulation électrique sur cette région cérébrales et leur impact sur la mémoire". - **Déborah Boyenval**, Team: MTV **(post-doctorante)** **Titre :** "Modelling the loss of differentiation potential during hematopoiesis through Boolean model inference" - **Gustavo Magaña López**, Team :MTV **Titre :** "Analysis of Boolean Network Ensembles via Probabilistic Context-Free Grammars and Vector Embeddings" - **Theo Roncalli**, Team: MTV **Titre :** "Boolean Networks Inference for Explaining Treatment-Resistant Leukemia" ***Département Image et Son (I&S)*** - **Colleen Beaumard**, Team TAD **Titre :** "Détecter la somnolence à partir de la parole avec l'analyse des phonèmes". - **Idris Dulau**, Team TAD **Titre :** "Characterization of the retinal vascular network" - **Kylian Desier**, Team TAD **Titre :** "Segmentation and classification of gliomas" - **Edern Le Bot:**, Team TAD **Titre :** "Segmentation holistique des IRM cérébrales par intelligence artificielle" - **Sebastien Richard**, Team TAD **Titre :** "Cross-modality Self-Supervised Learning for computer-assisted diagnosis of Age-related Macular Degeneration " - **Fabien Pelletier**, Team TAD **Titre :** "Deep Learning for Brain MRI Harmonization" - **David Trocellier**, Team MANIOC **Titre :** " Amélioration de la motricité post AVC en utilisant des neurotechnologies non invasives et de l'intelligence artificielle" ## 12h30 : Buffet