A 1,968-node coupled ring oscillator circuit for combinatorial optimization problem solving

Moy, William; Ahmed, Ibrahim; Chiu, Po-Wei; Moy, J.; Sapatnekar, Sachin S.; Kim, Chris H.

doi:10.1038/s41928-022-00749-3

Cited by 53 publications

(47 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Compared to our time-multiplexed experimental system in Fig. 1 , time-continuous analog Ising machines implement the Ising model in a fully analog way 14 , 16 , 37 , 38 . Here, spin state generation and spin coupling are performed in parallel, e.g.…”

Section: Resultsmentioning

confidence: 99%

“…Our method can easily be applied to existing large-scale analog Ising machines, e.g. based on optical parametric or electronic oscillators 16 , 38 , to extend their range of applications. By leveraging the inherently high analog bandwidth of such systems, this can lead to a generation of highly efficient statistical samplers that operate orders-of-magnitude faster compared to digital computers.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Noise-injected analog Ising machines enable ultrafast statistical sampling and machine learning

et al. 2022

View full text Add to dashboard Cite

Ising machines are a promising non-von-Neumann computational concept for neural network training and combinatorial optimization. However, while various neural networks can be implemented with Ising machines, their inability to perform fast statistical sampling makes them inefficient for training neural networks compared to digital computers. Here, we introduce a universal concept to achieve ultrafast statistical sampling with analog Ising machines by injecting noise. With an opto-electronic Ising machine, we experimentally demonstrate that this can be used for accurate sampling of Boltzmann distributions and for unsupervised training of neural networks, with equal accuracy as software-based training. Through simulations, we find that Ising machines can perform statistical sampling orders-of-magnitudes faster than software-based methods. This enables the use of Ising machines beyond combinatorial optimization and makes them into efficient tools for machine learning and other applications.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

Noise-injected analog Ising machines enable ultrafast statistical sampling and machine learning

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Recently, Moy et al reported a 1968-node coupled King’s graph ring oscillator–based Ising machine with five-level coupling strengths [ 44 ]. Here, the programmability was implemented by the transmission gates.…”

Section: Experimental Demonstrationsmentioning

confidence: 99%

“… ( a ) Ring oscillator network coupled through inverters [ 3 ]; ( b ) same-node and cross-node coupling configuration for positive and negative coupling, transmission gates controlled five coupling strengths [ 44 ]. …”

Section: Figurementioning

confidence: 99%

Oscillator-Network-Based Ising Machine

et al. 2022

View full text Add to dashboard Cite

With the slowdown of Moore’s law, many emerging electronic devices and computing architectures have been proposed to sustain the performance advancement of computing. Among them, the Ising machine is a non-von-Neumann solver that has received wide attention in recent years. It is capable of solving intractable combinatorial optimization (CO) problems, which are difficult to be solve using conventional digital computers. In fact, many CO problems can be mapped to finding the corresponding ground states of Ising model. At present, Ising machine prototypes based on different physical principles, such as emerging memristive oscillators, have been demonstrated, among which the Ising Hamiltonian solver based on the coupled oscillator network simultaneously holds the advantages of room-temperature operation, compact footprint, low power consumption, and fast speed to solution. This paper comprehensively surveys the recent developments in this important field, including the types of oscillators, the implementation principle of the Ising model, and the solver’s performance. Finally, methods to further improve the performance have also been suggested.

show abstract

“…Other recent studies have utilized CMOS electronic relaxation oscillators and, through chip integration, engineered considerable system sizes 44,46,47 . Our implementation allows for the formulation of physically realistic equations of motion, and thus one emphasis of this paper is on the physics-based exploration of the energy minimizer as a dynamical-systems process.…”

mentioning

confidence: 99%

An Ising machine based on networks of subharmonic electrical resonators

et al. 2022

View full text Add to dashboard Cite

Combinatorial optimization problems are difficult to solve with conventional algorithms. Here we explore networks of nonlinear electronic oscillators evolving dynamically towards the solution to such problems. We show that when driven into subharmonic response, such oscillator networks can minimize the Ising Hamiltonian on non-trivial antiferromagnetically-coupled 3-regular graphs. In this context, the spin-up and spin-down states of the Ising machine are represented by the oscillators’ response at the even or odd driving cycles. Our experimental setting of driven nonlinear oscillators coupled via a programmable switch matrix leads to a unique energy minimizer when one exists, and probes frustration where appropriate. Theoretical modeling of the electronic oscillators and their couplings allows us to accurately reproduce the qualitative features of the experimental results and extends the results to larger graphs. This suggests the promise of this setup as a prototypical one for exploring the capabilities of such an unconventional computing platform.

show abstract

A 1,968-node coupled ring oscillator circuit for combinatorial optimization problem solving

Cited by 53 publications

References 21 publications

Noise-injected analog Ising machines enable ultrafast statistical sampling and machine learning

Noise-injected analog Ising machines enable ultrafast statistical sampling and machine learning

Oscillator-Network-Based Ising Machine

An Ising machine based on networks of subharmonic electrical resonators

Contact Info

Product

Resources

About