268th RQC Seminar

• 講演者

  Mr. Lorenzo Fioroni
  ( The Swiss Federal Institute of Technology in Lausanne (EPFL) )

• 日程

  2026年4月21日(火), 16:00 - 17:00(4:00 p.m.-5:00 p.m.)

• 開催場所

  ハイブリッド（Zoom,
  345-347 Seminar Room, 3F, Main Research Building, Wako Campus / 和光地区 研究本館3階 セミナー室 (345-347) (C01))

• 講演タイトル

  Entanglement-assisted variational algorithm for discrete optimization problems

• お問合せ

  norilab_rqc_assist[at]ml.riken.jp

講演概要
From fundamental sciences to economics and industry, discrete optimization problems are ubiquitous across all fields. Yet, their complexity often renders exact solutions computationally intractable, necessitating the use of approximate methods. Among these, heuristic algorithms inspired by classical, and more recently quantum physics are widely employed. Quantum annealing holds promise over classical approaches by framing the problem as finding the ground state of a many-body spin system via adiabatic state preparation. Although still debated, it is believed that entanglement formation during the annealing process may lead to a speedup over classical methods.
We present a novel classical heuristic that efficiently simulates quantum annealing using a parameterized Ansatz of Generalized Coherent States. This framework allows for computation of energy and gradients with O(N^3) complexity, permitting efficient gradient descent optimization of large problems with over 2000 spins. At the same time, these states capture non-trivial entanglement structures, which are crucial for the effectiveness of quantum annealing.
We analyze our method by comparing solution quality and runtime to those of other popular heuristics. Our findings suggest that our method offers a scalable approach to leveraging quantum effects for complex optimization problems, potentially surpassing conventional heuristics in large-scale applications.