Front cover of SIAM News
Based on our PNAS paper ‘The difficulty of computing stable and accurate neural networks: On the barriers of deep learning and Smale’s 18th problem’ we have w...
Welcome
You have reached the webpage of Vegard Antun, a postdoctoral fellow in mathematics at the University of Oslo. Here you can find information about my research interests, publications, an overview of my code and contact details.
Research interests
Broadly speaking, I'm interested in applied mathematics, with a particular focus on inverse problems and imaging. In particular, I am interested in methods for recovery of signals (vectors, images, or functions) from a limited set of measurements. The study of such methods uses tools from many areas of mathematics including compressive sensing, approximation theory, mathematical signal processing as well as machine learning techniques such as deep learning. During the last few years, methods based on learning have started to outperform many standard methods, not only in computational imaging, but in scientific computing in general. Yet, many of these new approaches seem to have a common Achilles heel: stability. Much of my research focuses on the stability and accuracy of methods for inverse problems.
Additional projects
- Participating in the Z-Inspection: A process to assess trustworthy AI as a domain expert of two best practice usescases of AI in healthcare.
- I have contributed to the book Compressive Imaging: Structure, Sampling, Learning (Cambridge University Press) by B. Adcock and A. C. Hansen, by writing the code behind more than 350 images.
Background
- 2020-Present: Postdoctoral fellow - University of Oslo.
- 2016-2020: PhD candidate - University of Oslo. Supervised by Øyvind Ryan and Anders C. Hansen.
- 2014-2016: Master's degree - University of Oslo.
- 2013-2014: Military service - Military Police - Sessvolmoen/Kirkenes.
- 2010-2013: Bachelor's degree - University of Oslo.
Recent posts
IEEE Spectrum writes about our work!
Our PNAS paper ‘The difficulty of computing stable and accurate neural networks: On the barriers of deep learning and Smale’s 18th problem’ has been picked up...
PNAS publishes our paper on the barriers of deep learning
Our paper ‘The difficulty of computing stable and accurate neural networks: On the barriers of deep learning and Smale’s 18th problem’ has been published in t...
Launching a new study program
I’m proud to announce that we are launching the new study program “Fra data til innsikt” (From data to insight) at the University of Oslo. This is a study pr...
Welcome to Johan
Welcome to our new PhD student Johan S. Wind. Johan has been hired on the project “On the Barriers of AI in Scientific Computing” founded by dScience. I will...