Events Calendar
Variational Quantum Algorithms: An Overview
Thursday January 23, 2020
3:30 pm
Tweet |
Presenter: | Patrick Coles, Los Alamos National Lab |
---|---|---|
Series: | CQuIC Seminars | |
Abstract: | Variational Quantum Algorithms (VQAs) can be viewed as quantum generalizations of neural networks, where one evaluates the cost function with an efficient quantum circuit. Hence, analogous to neural networks, VQAs allow for task-oriented programming of quantum computers. VQAs have the additional benefit of error mitigation since they keep the circuit depth short, which is important for noisy quantum computers. In this talk, I will ambitiously aim to give you an overview of the field of VQAs, which is currently one of the hottest subfields of quantum computing. In particular, I will highlight three key topics: (1) New Applications, (2) Noise Resilience, and (3) Barren Plateaus. It is well known that the variational quantum eigensolver (the most famous VQA) is useful for elucidating the electronic structure of molecules for quantum chemistry. However, our Los Alamos group and others have recently presented VQAs for quantum compiling, state diagonalization, solving linear systems, and dynamical quantum simulation. While these new applications are exciting, perhaps even more exciting is our recent rigorously-proven result that certain VQAs exhibit resilience to noise, where noise does not affect the optimal parameters that one learns. Finally, I will present one of the most fascinating phenomena of quantum neural networks, which are exponentially vanishing gradients ("barren plateaus"), and I will discuss their impact on the scalability of VQAs. | |
Host: | Ivan Deutsch | |
Location: | PAIS-2540, PAIS | |