PhD Fellow at Integreat and the Department of Informatics, University of Oslo
Academic interests
I work in the intersection of theoretical machine learning and statistics, with a focus on understanding how complex models encode and learn data. I am interested in information theory, sufficient statistics theory, and deep learning. I am also interested in the integration of pre-existing knowledge into machine learning models.
Background
Bachelor's degree in mathematics and economics, and a master's in computational science with mathematical applications, both from the University of Oslo. The focus of my master’s thesis was physics-informed machine learning applied to dynamical systems. In the past years, I have worked as a research assistant for the Norwegian Computing Centre, SINTEF, and Norges Bank.
PhD project
“Embedded Sufficient Statistics”
This project aims to explore how layers in neural networks summarize and encode input data distributions. By combining concepts from theoretical machine learning and statistics, we will develop new methods to investigate how these layers compress and specialize information. The goal is to achieve near-optimal data encoding, addressing a long-standing challenge in machine learning related to the efficient representation of data.
Supervisors
- Ali Ramezani-Kebrya, University of Oslo, Department of Informatics
- Arnoldo Frigessi, University of Oslo, Department of Mathematics
- Sylvia Richardson, University of Cambridge, MRC Biostatistics Unit