Arjan Cornelissen wins ASML graduation price for mathematics
Arjan Cornelissen was nominated to the KHMW (Royal Dutch Society of Sciences) and has won the ASML graduation price for mathematics. The price consists of a sum of €5.000,- euro and will be awarded on 26 November in the Hodshonhuis in Haarlem. This national price for the best thesis in mathematics is awarded annually by the KHMW.
Arjan was guided within DIAM by Martijn Caspers and Jan van Neerven. The thesis is entitled: Quantum gradient estimation and its application to quantum reinforcement learning. The title shows that this project fits in with the (high-quality) initiatives that are currently emerging between the Analysis section and the Quantum world.
The thesis deals with quantum computing and the complexity of quantum algorithms. This is a relatively young field of study that deals with (optimization of) algorithms that can be executed on a quantum computer. In Arjan's thesis different techniques from mathematics and physics come together: quantum mechanics, complexity and functional analysis. Arjan has looked at quantum algorithms that can efficiently determine the gradient of a function. Such gradients have important applications in machine learning and gradient descent algorithms. An important and surprising result in Arjan's thesis is that he shows that for functions with a certain smoothness (so-called Gevery classes) there are algorithms to determine their gradient quickly and he gives a lower limit on the speed of an optimal algorithm.
More information about the thesis https://repository.tudelft.nl/islandora/object/uuid:26fe945f-f02e-4ef7-bdcb-0a2369eb867e?collection=education