Research

My research is in data science and AI for scientific applications. It is by nature interdisciplinary. It focuses on developping computer vision, machine learning, and deep learning methods for analysing data from other sciences. The application domains that I consider fit into two broad categories: physical sciences, and human sciences. There is a special focus on the astrophysics and medical fields. My research is both applied and theoretical. It aims to develop new tools to support research in the (physics and human science) application domains. It also aims to further develop AI methods, motivated by new challenges from the scientific data. One strong challenge, that is common to all considered application domains, is the need for interpretable methods and results. This is currently being addressed by the development of domain informed models, where domain knowledge is integrated into the design of learning methods.

Research themes

• Computer vision, machine learning, deep learning
• Physics & astrophysics applications
• Human sciences & medical applications
• Domain informed learning: interpretable AI through the integration of domain knowledge into the design of learning methods

Ongoing projects

Physics applications

Solar Physics

• PRESAGE: PREdicting Solar Activity using machine learning on heteroGEneous data
• ANR JCJC 2021-2025
• Collaboration with Observatoire de Paris

• Detection and characterisation of type II solar radio bursts
  • ERUPTION: dEtection de suRsaUts radio solaires par deeP learning avec inTegratION de principes physiques, CNRS-INS2I “Volet Emergence” 2020
  • MRes supervision, Joseph Jenkins, 2018-2019
  • Collaboration with Observatoire de Paris and Space Environment Laboratory at Tokyo

Planetology

• ATLAS: computer vision and deep learning methods for Analysing mulTimodaL mArs imageS towards a global interpretation of martian terrain properties
• • Région Sud “Emploi Jeunes Doctorants” 2023-2026
  • Co-funding by private company ACRI-ST
• Collaboration with Institut de Planétologie et d'Astrophysique de Grenoble (IPAG)
• PhD co-supervision: Loïs Brun since 2023, Kévin Kheng 2018-2022

Galaxy evolution

• SIDEREAL: phySics-Informed Deep nEuRal nEtwork for characterising gALaxy morphology
• • CNRS-MITI “Sciences pour l’IA, l’IA pour les sciences” 2022-2023
  • CNRS-MITI 80Prime 2023
• Collaboration with Observatoire de Strasbourg and MATLAS
• PhD co-supervision: Renaud Vancoellié since 2023, Felix Richards 2017-2022

Oceanography

• Prediction of Lagrangian drift at sea
  • PhD co-supervision, Joseph Jenkins, since 2020
  • Collaboration with MIO laboratory, OceanNext, Datlas
• Mapping of ocean dynamics from satellite images
  • PhD co-supervision, Laura Gibbs, since 2018
  • Collaboration with Department of Oceanography at University of Bristol

Human sciences applications

Medical

• MyOCARdE: Mesure de la fOnction Cardiaque basée sur l’AppREntissage - A new learning-based measure for heart function
• Région PACA 2019-2022
• Collaboration with Bristol Heart Institute

Linguistic

• DRAGON-S: Developing Resistance Against Grooming Online - Spot and Shield
• UNICEF/EVAC 2020-2024
• Collaboration with Department of Linguistic at Swansea University & TARIAN-ROCU

Past projects

Physics applications

Chemistry

• Prediction of chemical properties and discovery of new crystal structures using deep learning on graphs
  • PhD supervision, Jay Morgan, 2018-2021
  • Collaboration with the Departments of Chemistry at Swansea University (UK) and at University of Rostock (Germany)

Human sciences applications

Medical

• Online quality assessment of human movements from Kinect skeleton data
• Project [SPHERE](https://research-information.bris.ac.uk/en/projects/sphere-epsrc-irc), computer vision workpackage

• IReSISD: Integrated Registration, Segmentation, and Interpolation for 3D/4D Sparse Data
• Doctoral school, University of Bristol
• Collaboration with Bristol Heart Institute