Thesis: Electron Energy Reconstruction
In my bachelor thesis I attempt to replace the current Bremsadder function (responsible for reconstructing the energy lost through Bremsstrahlung) at LHC by using machine learning based techniques.
Supervised by Jacco de Vries
Improving SLAM using CNNs
Here CNNs are used to create a 3D representation of the environment from a stereo camera. This is done in an effort to improve upon the previous project by allowing the drone to avoid objects.
Led by Szymon Fonau & Me
Autonomous Drone Tracking
Computer Vision is used to track humans and animals using a drone. This drone is able to autonomously follow moving targets at a constant distance.
Led by Szymon Fonau & Me
Particle Identification at LHCb
A deep neural network was built to test whether an addition measurement of a parameter (velocity) added any useful information for determining the difference between Kaons and Pions. This could be confirmed with significantly increased accuracy for the measurement of slow moving particles.
Supervised by Jacco de Vries
Search Strategies by Artificially Intelligent Agents
The search behaviour of simple larvae was recreated using a genetically evolving deep neural network. Certain movement patterns found in nature naturally reoccured in the trained neural networks and gave indications for the sensitivity of the larvae's senses.
Supervised by Chris Pawley