Mohammad Sadegh Salehi

PhD student in Applied Mathematics (SAMBa CDT)
Department of Mathematical Sciences
University of Bath

Email & Social media: researchgate
Office Location: 4W 1.17


About me

    I am a PhD student at the Department of Mathematical Sciences, University of Bath and a member of the Numerical Analysis group and the Bath Imaging Group, under the supervision of Dr. Matthias J. Ehrhardt & Dr. Subhadip Mukherjee. I am interested in Numerical Optimisation, Machine Learning, Deep Learning, Inverse Problems, and Image Analysis. My current research lies in the field of "Bilevel Learning". In particular, I am working on gradient-based and derivative-free algorithms for Bilevel Learning, as well as considering Unrolled Approach and input-convex Neural Networks as the regulariser. The applications I consider now include undersampled MRI reconstruction; image denoising, deblurring and segmentation; and clustering.

Research Projects

  • Gradient Free Training of Neural Networks (ADMM Approach)
    Matthias Ehrhardt, Mathematics of Deep Learning Reading course, October 2021- January 2022
  • Stochastic Quasi Newton Methods vs Stochastic Gradient
    Matthias Ehrhardt & Clarice Poon, Large-scale Optimisation for Machine Learning Reading course, February 2022- June 2022
  • Interdisciplinary Research Project
    Alex Cox, Portfolios Design for Clinical Trials, October 2021- April 2022
  • Master's Thesis
    Antonin Chambolle, "Linear and non-linear acceleration methods for alternating projections and Dykstra’s algorithm for projecting on an intersection of convex sets and an optimal transport problem", March 2021- August 2021
© 2021-2022 Mohammad Sadegh Salehi - - Last Update: 07/2022