Online resume under construction.
I am a master student at Chalmers School of Technology, specializing in machine learning and algorithms. I am currently writing my thesis on the use of reinforcement learning within the field of radar resource management.
I will be graduating in June of 2021, and am interested in pursuing a career within the field of Machine- and Deep Learning. I am especially interested in increasing the accessibility of machine learning to the general public and applications within medicine and Medtech.
If you have any questions or want to get in touch you can reach me at my email:
E-mail: axel.nathanson(at)gmail.com
Linkedin
I am currently writing my Master thesis for Saab on utilizing Deep Reinforcement Learning to optimise the management of radar resources while performing surveillance and tracking. The project includes the simulation and tracking of the targets as well as implementation of DRL - architectures.
The system is modelled as a POMDP why recurrent architectures and the PPO-algorithm will be the initial focus. However, since the number of targets within the surveillance area varies over time techniques for handling varying input-size in Deep Learning will also be explored.
I am writing my thesis alone but with the guidance of my supervisor Adam Andersson, whose current research topic concerns the use of deep learning to obtain accelerated and scalable algorithms for computationally challenging problems.
I created this webpage to be able to showcase some of the projects I have carried out.
One of the courses that have inspired me the most was Machine Learning for Graphs and Sequential Data that I took during my exchange at the TUM. The course covered a wide variety of topics, but I found the topics of Generative models and Robustness especially interesting.
As an exercise in implementing models from scientific papers, I have implemented two generative models since the course ended which can be found at:
After I attended an introductory course in Deep Learning, I wanted to create a model and all functions needed from scratch as an exercise. The course focused on Computer Vision, why I chose the most classic Computer Vision-task possible:
The model created uses the most basic building blocks in Deep Learning for Computer Vision like CNN:s, Batch Norm-layers, Dropout-layers, Data Augmentation.
Engineering mathematics and computational science, 2019-2021
My master programme extended my knowledge of mathematics and statistics while giving me the computational tools to use them. I have focused my studies within the fields of algorithms, statistics and machine learning.
Faculty of Mathematics, 2019-2020
I studied the first year of my Master degree at the TUM. There I read courses in computational statistics and machine- and deep learning, igniting my interest in the field.
Engineering Physics, 2016-2019
My undergraduate physics programme covered a broad range of subjects in physics, mathematics and engineering. The very diverse syllabus covers courses like quantum physics, complex analysis as well as algorithms and statistics.
For my bachelor thesis, I worked with the department of physics on Simulating Many-Particle Systems on an Emulated Quantum Computer (in Swedish). My main responsibility during the project was optimization.
Law School, 2014-2015
I studied for three semesters at one of the top law school in Sweden. During my studies, I realised that I was more interested in mathematics and physics, but did take courses in:
Project, Nov 2020 - Jan 2021
I was part of a group of students who worked with IRLAB and Smartr to model an unknown filtering algorithm. The algorithm is implemented in an old analyser-unit used to record the trajectories of rats during experimental trials. As part of the study, both known filtering algorithms like Kalman-filters and Splining-algorithms were evaluated as well as new ones created. As a result of the study, we could rule out several classic filtering algorithms while achieving the best results with the algorithms designed during the study. My main responsibility was the statistical analysis of the trajectories, providing the basis for the algorithm design.
Link to project description.
Student Intern, Jun 2018 - Sep 2019
During my time with the Vehicle Homologation and Compliance team, I was responsible for the migration of all legal requirements into two new databases. One database for the internal review of the legal requirements, and one used to communicate them during projects. I also assisted in the mapping of new legal areas.