About

LinkedIn   GitHub

I’m working at GoPro on deep learning and embedded image processing. I did my PhD at the Toulouse Mathematics Institute under the supervision of Pierre Weiss and Frédéric de Gournay, working primarly on inverse problems, optimization and neural networks.

I received a Master of Science in Fundamental Mathematics and Applications (PDE, numerical analysis and optimization track) from Université Paul Sabatier in 2019. I also graduated from INSA Toulouse in Applied Mathematics (numerical modelling and image processing) in 2019.

Research

Preprints

• Adaptive scaling of the learning rate by second order automatic differentiation
  F. de Gournay & A. Gossard. PDF

Journal papers

• Training Adaptive Reconstruction Networks for Blind Inverse Problems
  A. Gossard & P. Weiss. SIAM Journal on Imaging Sciences. PDF

• Bayesian Optimization of Sampling Densities in MRI
  A. Gossard, F. de Gournay & P. Weiss., Journal of Machine Learning for Biomedical Imaging (MELBA), Volume 2 (2023). PDF

• Spurious minimizers in non uniform Fourier sampling optimization
  A. Gossard, F. de Gournay & P. Weiss, Inverse Problems (2022). PDF

Conference papers

• iTWIST, 2020, remote. Off-the-grid data-driven optimization of sampling schemes in MRI. With Frédéric de Gournay & Pierre Weiss. PDF

Oral presentations

• Applied Inverse Problems (AIP) Conference, Göttingen (Germany), September 2023, Learning Fourier sampling schemes for MRI by density optimization.
• Torus AI Tasting seminar, June 2023, Training adaptive reconstruction networks for blind inverse problems.
• PhD defence, Toulouse, December 2022, Learning strategies for computational MRI.
• Curves & Surfaces 2022, Arcachon, June 2022, Spurious minimizers in non uniform Fourier sampling optimization.
• SIAM Imaging Science 2022, remote, March 2022, Learning to sample and reconstruct: computational issues and tricks.
• Séminaire doctorant de l’IMT, Toulouse, March 2022, Globalization approach for sampling scheme optimization in MRI.
• EUROPT 2021, remote, July 2021, Continuous optimization of Fourier sampling schemes for MRI.
• iTWIST 2020, remote, December 2020, Off-the-grid data-driven optimization of sampling schemes in MRI.

Volunteer experience

Involved in the university’s fablab Fabric’INSA since 2017.

Former president of the Gamma association

Recently I co-founded the Opteromys association. Opteromys is an association federating people interested in the development of low carbon emitting technologies for transportation.

Contact

@lb@n.p@ul.goss@rd@gm@il.com (replace ‘@’ by ‘a’ at relevant places)

GPG fingerprint: 3D13 BC84 EB50 12F1 B5E1 3F05 5261 7F1C F64F F7BC