Unsupervised learning with Rigetti 19Q

Qu&Co comments on this publication:

Practical applications for current noise and small quantum-computing hardware, has focused mostly on short-depth parameterized quantum circuits used as a subroutine embedded in a larger classical optimization loop. In this paper, Otterbach et al. describe experiments with unsupervised machine learning (specifically clustering), which they translate into a combinatorial optimization problem solved by the quantum approximate optimization algorithm (QAOA) running on the Rigetti 19Q (a 19 qubit gate-based processor). They show that their implementation finds optimal solution for this task even with relatively noisy gates.