Hana Sebia ☕️
Hana Sebia

PhD candidate in Machine Learning

About Me

Ph.D. candidate at Inria, the French national research institute for digital science and technology. My research focuses on machine learning methods for tabular data and medical imaging. I work on developing interpretable and efficient models to extract meaningful representations from high-dimensional data. My interests include temporal modeling and tensor factorization for structured data, as well as deep learning approaches for image segmentation, generation, and modality transfer.

Download CV
Interests
  • Deep Learning
  • Image Segmentation
  • Tensor Decomposition
  • Image Quality Transfer
  • Healthcare Data Analytics
  • Computational Phenotyping
Education
  • Ph.D. in Machine Learning

    Inria, France

  • Master's degree in Computer Science, Data Science option

    Claude Bernard Lyon 1 University

  • Bachelor's degree in Computer Science, Math option

    Claude Bernard Lyon 1 University

Experience

  1. Visiting Researcher

    Centre of Medical Image Computing, University College London

    London, UK

    Collaborators : Daniel Alexander, Joseph Jacob

    • Vascular segmentation of HiP-CT brain images
    • Synthesize Utrasound Localisation Microscopy (ULM) images from Functional Ultrasound (fUS) using deep generative models
  2. PhD Candidate

    AIstroSight, Inria

    Lyon, France

    Supervisors : Thomas Guyet, Hugues Berry

    Collaborators : Benjamin Vidal, Etienne Audureau

    • [fUS-ULM] : Functional Ultrasound Imaging captures rapid hemodynamic changes but offers limited spatial resolution. Ultra Localized Microscopy provides high-resolution images but suffers from long acquisition times. Combining fUS and ULM could produce detailed images with both high resolution and fast frame rates.

    • [AI-RACLES] : A chair funded by Inria-APHP-CS aiming to develop new artificial intelligence techniques to better exploit the greater Paris university Hospital (APHP) data lake. The context of this internship is to investigate how to support the evaluation of health care pathways.

  3. Healthcare Data Analyst

    LIRIS

    Lyon, France

    Supervisor : Hacid Mohand Saïd

    Collaborators : Delphine Maucort-Boulch, François Talbot

    • [KANOPEE]: An application offering clinical identification and advice by a virtual companion to limit sleep problems and addictive behaviors, early markers of anxiety, stress and depression linked to the COVID-19 crisis.

    • [QUALITOP]: European project aiming to develop a smart digital platform using big data analysis to monitor health status and quality of life of cancer patients given immunotherapy.

Education

  1. Ph.D. in Machine Learning

    Inria, France
  2. Master's degree in Computer Science, Data Science option

    Claude Bernard Lyon 1 University
  3. Bachelor's degree in Computer Science, Math option

    Claude Bernard Lyon 1 University
Latest Publications
All Publications
(2025). Spatiotemporal pattern extraction from functional neuroimaging. Conférence en Apprentissage automatique (CAp), Dijon, France.
(2025). Vascular segmentation of functional ultrasound images using deep learning. Computers in Biology and Medicine.
(2025). Vascular Segmentation of fUS Images using Deep Learning. Intelligence Artificielle en Imagerie Biomédicale (IABM), Nice, France.
(2024). SWoTTeD: an extension of tensor decomposition to temporal phenotyping. Machine Learning.
(2022). Une extension de la décomposition tensorielle au phénotypage temporel. EGC 2023-23ème Conférence Francophone sur Extraction et Gestion des Connaissances.
Recent & Upcoming Talks