Experimental Investigation of the Dynamics of Coupled Oscillators as Ising Machines

Bashar, Mohammad Khairul; Mallick, Antik; Shukla, Nikhil

doi:10.1109/access.2021.3124808

Cited by 5 publications

(8 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We observed a similar behavior even for graph instances with three optimal states. This suggests there are optimal critical points that are unstable (such as saddle points) as recently highlighted by Bashar et al [47].…”

Section: Weighted Max-cut On Pcbmentioning

confidence: 71%

“…Beyond-CMOS devices are promising as they generally allow a compact oscillator design using a single device that could be scaled down to the nanoscale. Nevertheless, CMOS-based ONNs benefit from the mature CMOS technology which enables rapid ONN development and facilitates its co-integration with conventional digital circuits [5,12,20,21]. In this work, we focus on ONNs that compute in phase domain, i.e.…”

Section: State-of-the-art Onnsmentioning

confidence: 99%

“…For solving COPs, a general approach is to map the input graph to the ONN where vertices are oscillators, and edges are synapses. Some architectures such as [10,20] are dedicated to finding the maximum cut of a graph with weights of the same sign, as the synapses only implement negative weights. The main drawback is that both coupling capacitors and resistors are required to program negative and positive weights, respectively.…”

Section: State-of-the-art Onnsmentioning

confidence: 99%

“…However, due to the high dimension of the problem relaxation (R N ), GW is costly for large instances [38,39] and alternative approaches using physical systems such as quantum annealers [40], coherent Ising machines [39], memristors [41] or coupled oscillators are being investigated [9,10,12,20].…”

Section: Skonn For Solving the Np-hard Max-cut Problemmentioning

confidence: 99%

“…In this work, we introduce a new mixed-signal ONN architecture, named Saturated Kuramoto ONN (SKONN) that leverages both analog and digital domains to satisfy the three ONN design criteria. SKONN takes inspiration from the state-of-the-art analog ONN architectures for which the dynamics evolve naturally in continuous time and can easily be described by phase models like Kuramoto's or Izhikevich's [14], thus Goto [3] Jackson et al [5] Wang et al [18] Chou et al [19] Bashar et al [20] Dutta et al [10] Moy et al [12] Graber et al [21] This work facilitating the exploration of potential applications. SKONN's main novelty consists in setting the computation and propagation in the analog and digital domains, respectively.…”

Section: A Mixed-signal Approachmentioning

confidence: 99%

See 4 more Smart Citations

A mixed-signal oscillatory neural network for scalable analog computations in phase domain

Delacour¹,

Carapezzi²,

Boschetto³

et al. 2023

Neuromorph. Comput. Eng.

View full text Add to dashboard Cite

Digital electronics based on von Neumann’s architecture is reaching its limits to solve large-scale problems essentially due to the memory fetching. Instead, recent efforts to bring the memory near the computation have enabled highly parallel computations at low energy costs. Oscillatory Neural Network (ONN) is one example of in-memory analog computing paradigm consisting of coupledoscillating neurons. When implemented in hardware, ONNs naturally perform gradient descent of an energy landscape which makes them particularly suited for solving optimization problems. Although the ONN computational capability and its link with the Ising model are known for decades, implementing a large-scale ONN remains difficult. Beyond the oscillators’ variations, there are still design challenges such as having compact, programmable synapses and a modular architecture for solving large problem instances.In this paper, we propose a mixed-signal architecture named Saturated Kuramoto ONN (SKONN) that leverages both analog and digital domains for efficient ONN hardware implementation. SKONN computes in the analog phase domain while propagating the information digitally to facilitate scaling up the ONN size. SKONN’s separation between computation and propagation enhances the robustness and enables a feed-forward phase propagation that is showcased for the first time. Moreover, the SKONN architecture leads to unique binarizing dynamics that are particularly suitable for solving NP-hard combinatorial optimization problems such as finding the Weighted Max-cut of a graph. We find that SKONN’s accuracy is as good as the Goemans-Williamson 0.878-approximation algorithm for Max-cut; whereas SKONN’s computation time only grows logarithmically. We report on Weighted Max-cut experiments using a 9-neuron SKONN proof-of-concept on PCB. Finally, we present a low-power 16-neuron SKONN integrated circuit and illustrate SKONN’s feed-forward ability while computing the XOR function.

show abstract

Section: Weighted Max-cut On Pcbmentioning

confidence: 71%

Section: State-of-the-art Onnsmentioning

confidence: 99%

Section: State-of-the-art Onnsmentioning

confidence: 99%

Section: Skonn For Solving the Np-hard Max-cut Problemmentioning

confidence: 99%

Section: A Mixed-signal Approachmentioning

confidence: 99%

See 3 more Smart Citations

A mixed-signal oscillatory neural network for scalable analog computations in phase domain

Delacour¹,

Carapezzi²,

Boschetto³

et al. 2023

Neuromorph. Comput. Eng.

View full text Add to dashboard Cite

show abstract

Highly reconfigurable oscillator-based Ising Machine through quasiperiodic modulation of coupling strength

Albertsson

Rusu

2023

Sci Rep

View full text Add to dashboard Cite

Ising Machines (IMs) have the potential to outperform conventional Von-Neuman architectures in notoriously difficult optimization problems. Various IM implementations have been proposed based on quantum, optical, digital and analog CMOS, as well as emerging technologies. Networks of coupled electronic oscillators have recently been shown to exhibit characteristics required for implementing IMs. However, for this approach to successfully solve complex optimization problems, a highly reconfigurable implementation is needed. In this work, the possibility of implementing highly reconfigurable oscillator-based IMs is explored. An implementation based on quasiperiodically modulated coupling strength through a common medium is proposed and its potential is demonstrated through numerical simulations. Moreover, a proof-of-concept implementation based on CMOS coupled ring oscillators is proposed and its functionality is demonstrated. Simulation results show that our proposed architecture can consistently find the Max-Cut solution and demonstrate the potential to greatly simplify the physical implementation of highly reconfigurable oscillator-based IMs.

show abstract

Operating Coupled VO₂-Based Oscillators for Solving Ising Models

Avedillo,

Través,

Delacour

et al. 2023

IEEE J. Emerg. Sel. Topics Circuits Syst.

View full text Add to dashboard Cite

Experimental Investigation of the Dynamics of Coupled Oscillators as Ising Machines

Cited by 5 publications

References 42 publications

A mixed-signal oscillatory neural network for scalable analog computations in phase domain

A mixed-signal oscillatory neural network for scalable analog computations in phase domain

Highly reconfigurable oscillator-based Ising Machine through quasiperiodic modulation of coupling strength

Operating Coupled VO₂-Based Oscillators for Solving Ising Models

Contact Info

Product

Resources

About