Speaker
Prof.
Oleksandr Kyriienko
Description
In the talk I will describe recent advances in building quantum models that can learn on tasks with symmetries and laws of physics in the form of differential equations. First, I will introduce approaches based on differential constraints and discuss their applications in generative modelling. Second, I will discuss the concept of quantum data and present an analysis for embeddings motivated by physical processes. Finally, I will show that embedding symmetries into quantum machine learning models can help discovering protocols able to solve Simon's and forrelation problems with excellent generalization.
