Large-scale Ising spin network based on degenerate optical parametric oscillators

By T. Inagaki, K. Inaba, Ryan Hamerly, Kyo Inoue, Yoshihisa Yamamoto, Hiroki Takesue. Nature Photonics, 2016

223 citations recorded by the source corpus.

Solving combinatorial optimization problems is becoming increasingly important in modern society, where the analysis and optimization of unprecedentedly complex systems are required. Many such problems can be mapped onto the ground-state-search problem of the Ising Hamiltonian, and simulating the Ising spins with physical systems is now emerging as a promising approach for tackling such problems. Here, we report a large-scale network of artificial spins based on degenerate optical parametric oscillators (DOPOs), paving the way towards a photonic Ising machine capable of solving difficult combinatorial optimization problems. We generate >10,000 time-division-multiplexed DOPOs using dual-pump four-wave mixing in a highly nonlinear fibre placed in a cavity. Using those DOPOs, a one-dimensional Ising model is simulated by introducing nearest-neighbour optical coupling. We observe the formation of spin domains and find that the domain size diverges near the DOPO threshold, which suggests that the DOPO network can simulate the behaviour of low-temperature Ising spins. More than 10,000 time-division-multiplexed degenerate parametric oscillators are generated using phase-sensitive amplification in a nonlinear optical fibre. They can be used to simulate a coherent Ising machine that could solve difficult computing problems.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/a914c73255e15b15484838ef9f71228f484cc48a

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Experimental investigation of performance differences between coherent Ising machines and a quantum annealer (2018, background): “Problems with sparse connectivity, such as one-dimensional chains [compare (51) and (52)] and MAX-CUT on cubic graphs (Fig.”
Ising machines as hardware solvers of combinatorial optimization problems (2022, background): “However, experimental realizations of CIMs thus far 35,115,116,121,125 have been in regimes of high photon loss, where purely classical models can accurately describe the pertinent dynamics of the systems.”
Large-Scale Optical Neural Networks based on Photoelectric Multiplication (2018, methodology): “This encoding, reminiscent of the Coherent Ising Machine [28, 29, 30], contrasts with other approaches used for neural networks, which encode inputs in separate spatial channels [23, 24, 22].”
Stochastic p-bits for Invertible Logic (2016, background): “(1) in order to emphasize the generality of the concept of p-bits, which need not necessarily be nanomagnet based [17,18].”
A fully programmable 100-spin coherent Ising machine with all-to-all connections (2016, background): “Initial experiments with optical-delay-line-coupled OPOs have demonstrated that they can find solutions to three specific Ising problems: a complete graph with N = 4 spins (18); a MöbiusLadder graph with N = 16 spins (24); or a 1D Ising spin chain at large scale (25).”