Pubblicata Giovedì 13 Giugno 2024 11:02
Seminario Dr Alberto Nocera - Computational supremacy in quantum simulation
19 June 2024,16:30- Physical Seminar Series Announcement!
The overarching goal of noisy intermediate-scale quantum era is the demonstration that Quantum Processing Units (QPU) can solve a problem of practical interest that is impossible for any existing classical computer.
A natural playground for such a demonstration is the simulation of coherent Schrödinger dynamics of a quantum-mechanical system, where the analogue quantum simulator closely mimics the quantum-mechanical system's physical microscopic degrees of freedom and their interactions. Focusing on the
non-equilibrium dynamics of a magnetic spin-glass system (in two, three, and infinite dimensions) quenched through a quantum phase transition, I show that the latest D-Wave superconducting quantum annealing processor faithfully simulates the quantum spin-glass dynamics in close agreement with solutions of the Schrödinger equation solved using the best known classical simulation method for the problem: time-dependent variational principle applied to Matrix Product States (MPS). I show that MPS simulations are key to establish computational supremacy of the D-Wave QPU: extrapolation of classical computational resources beyond the simulable
scale of a few hundred qubits, far below the maximum experimental scale, amount to a hypothetical runtime of millions of years on the fastest supercomputer in the world, with infeasible memory and energy requirements.
Finally, I show that verification of the D-Wave QPU beyond the classically tractable system sizes is established measuring critical exponents closely matching estimates from the expected universality class using Kibble-Zurek theory. References: A. D. King, A. Nocera, et al. arXiv:2403.00910 [quant-ph]
Venue: Classroom 32 C3 of the Department of Physics, Building 32 C 6th
Floor.
More info at the link
After the seminar a light refreshment will be served to favor discussions