Seminar on Quantum Reservior Computing

Johannes Nokkala  from the University of Turku, Finland.

Quantum reservoir computing is a paradigm for harnessing the dynamics of driven-dissipative quantum systems to solve machine learning tasks. It's attractive because it presents a remarkably direct path from hardware to applications combined with extremely fast and robust training. Still, in principle just a few qubits can do the job of hundreds of classical neurons. Furthermore, unlike its classical counterparts, quantum reservoir computing is also suitable for coherent sensing and processing of quantum signals. Since it's introduction in 2017, it has become clear that quantum reservoir computing can tolerate realistic experimental conditions, but reaching the classically hard regime under such conditions and harnessing it properly remain important challenges. I will present an introduction and an overview of this promising but still young field, focusing on research done in the University of Turku. I will conclude by presenting some open questions the field is currently trying to answer.