Phasecraft and Leeds researchers propose FRIO, a reverse-annealing algorithm for spin glasses
Researchers at Phasecraft and the University of Leeds describe Fast Reverse Iterative Optimisation (FRIO), a quantum algorithm for finding ground states of spin glass models such as the Sherrington-Kirkpatrick model. It uses continuous reverse quantum annealing across a transition into a many-body localised phase, then digitises the process into circuits with parameters set through thermalizing dynamics, aimed at near-term devices. In numerical tests, FRIO almost always outperformed the Quantum Approximate Optimisation Algorithm (QAOA) while requiring fewer steps, but only on small instances of up to 12 (the article does not say whether these are spins or qubits). The work is posted on arXiv under the title 'Non-adiabatic Optimisation via Localisation-Delocalisation Transitions' by Ryan Smith, Sabrina Yue Wang, Filippo Maria Gambetta and Raul A. Santos of Phasecraft; the article does not say it has been peer reviewed, so it should be treated as a preprint that is not yet peer reviewed.