Quantum Technology Initiative Journal Club

Thursday 15 Dec 2022, 16:00 → 17:00 Europe/Zurich

513/R-070 - Openlab Space (CERN)

Michele Grossi (CERN), Sofia Vallecorsa (CERN)

Weekly meeting to present recent quantum machine learning related papers

16:00 → 17:00 On the role of noise and optimizer in studying Barren Plateau effect in Variational Quantum Circuits

In the Noisy Intermediate-Scale Quantum era (NISQ), Variational Quantum Algorithms (VQAs) are a valu- able resource in the path towards quantum advantage. Since they rely on the optimization of cost functions C, dependent on parameterized quantum circuits, performance of the optimizer and the study of the op- timization landscapes are of crucial importance. In the first part of the thesis, the optimization strategy is considered, and a variant of natural gradient descent is generalized to the case of Variational Quantum Classifiers (VQCs) with batched inputs. Then the influence of noise in the appearance of the barren plateau phenomenon is analyzed using with numerical simulations. The effect of different types of quantum channels is separately studied to assess the dominant contribution. Finally, the problem of scaling the model size is tackled in two dimensions, by scaling both the number of qubit q and the number of layers L. By doing so, the reference plots present in the literature are extended to a newly proposed visualization in the two dimensional model’s configuration space.

Conveners: Giulio Crognaletti, Michele Grossi (CERN)

recording