Pith. sign in

REVIEW 3 major objections 5 minor 2 cited by

How to Incorporate Higher-order Interactions in Analog Ising Machines

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Pith's one-line read The paper argues that analog Ising machines should feed higher-order spin interactions back through the signs of the spins rather than their continuous amplitudes, and shows on uniform random 3-SAT benchmarks that this sign-based rule…

desk verdict A clean comparative benchmark shows spin-sign local fields win for higher-order terms in one analog IM dynamics; the generalization beyond that dynamics is asserted more strongly than tested. read the letter →

arxiv 2507.23621 v1 pith:SX7PTJJO submitted 2025-07-31 cond-mat.stat-mech math-phmath.MPnlin.AOnlin.CGphysics.app-ph

classification cond-mat.stat-mechmath-phmath.MPnlin.AOnlin.CGphysics.app-ph PACS 05.50.+q
keywords analogIsingmachineshigher-orderinteractions3-SATspin-signlocalfieldsamplitudeimbalancetime-to-solutionsuccessratesmoothapproximations
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper asks how an analog Ising machine---a device whose spins are continuous amplitudes rather than discrete $\pm 1$---should include three-spin terms when solving 3-SAT problems. The authors compare five ways of building the local field that drives each spin, including three amplitude-rescaling variants and a direct baseline. They find that replacing each spin by its sign in all interaction terms, rather than its raw amplitude, consistently solves more benchmark instances and finds solutions faster, and the advantage widens as problem size grows. They also show that a smooth approximation of the sign rule, which is needed for physical hardware, retains the benefit when the transition is sharp enough. If correct, this gives a simple design rule for higher-order analog Ising machines.

What carries the argument

The load-bearing object is the local-field choice in the spin-update equation $\frac{ds_i}{dt}=-s_i+\tanh(\alpha s_i+\beta I_i)$, specifically the spin-sign replacement of Eq. (9): $I_i = J_i^{(1)} + \sum_j J_{ij}^{(2)}\operatorname{sgn}(s_j) + \sum_{j<k} J_{ijk}^{(3)}\operatorname{sgn}(s_j)\operatorname{sgn}(s_k)$. Because each spin enters with fixed magnitude $\pm 1$, low-order terms cannot drown out high-order terms when amplitudes are near zero, preserving the relative weighting of interactions written into the coefficients. The baseline is the direct amplitude-based local field; the three rescaling methods multiply spin amplitudes by powers of the mean absolute spin value to realign the orders; the sign method is the only one that enforces binary contributions. The $\tanh$ nonlinearity keeps amplitudes in $[-1,1]$, making the sign well defined, and the smooth approximation $\tanh(\kappa s_i)$ stands in for the discontinuous sign in hardware.

What would settle it

Run the same five local-field formulations on the same family of uniform random 3-SAT instances but with a modified update rule, for instance adding momentum or chaotic amplitude control, using the same hyperparameter search; if one of the rescaling methods then achieves lower time-to-solution and higher success rate on a majority of instances, the paper's claim that sign-based interactions are the best way to incorporate higher-order terms would be limited to the original dynamics. Alternatively, a hardware device whose smooth sign approximation is capped below the steepness at which the simulation reaches parity would falsify the hardware-compatibility claim for that device.

Watch

Extended reading notes

Core claim

The paper's discovery is that the standard way of putting a higher-order Hamiltonian into an analog Ising machine is mis-scaled: if the binary spins $\sigma_i=\pm 1$ are relaxed to continuous amplitudes $s_i\in[-1,1]$, a local field built from the derivative of the energy makes linear terms dominate quadratic terms and quadratic terms dominate cubic terms whenever amplitudes are small. The paper's fix, Eq. (9), replaces each amplitude inside the local field by its sign, so every interaction contributes with unit magnitude: $I_i = J_i^{(1)} + \sum_j J_{ij}^{(2)}\operatorname{sgn}(s_j) + \sum_{j<k} J_{ijk}^{(3)}\operatorname{sgn}(s_j)\operatorname{sgn}(s_k)$. On uniform random 3-SAT instances ranging from 20 to 250 spins, this sign-based rule solves more instances within the time budget and reaches solutions faster than the baseline and all three rescaling methods, and the gap grows with problem size. A smooth approximation $\tanh(\kappa s_i)$ recovers the same performance once $\kappa$ is large enough, showing the rule is compatible with finite-steepness analog hardware. The authors present the result as a generalization of their earlier finding for external fields in quadratic cost functions: enforcing binary contributions in the local field preserves the discrete structure of the problem.

Load-bearing premise

The performance ranking was measured under one fixed choice of machine dynamics (a smooth saturating update without momentum and with a linear annealing schedule); a different dynamics could change which local-field method wins.

Editorial extensions

If this is right

  • Sign-based local fields should replace amplitude rescaling as the default recipe for cubic and higher-order terms in analog Ising machines; the paper finds them faster and more reliable across all tested problem sizes.
  • The advantage grows with problem size, so the rule matters most in the large-scale regime where Ising machines are meant to compete.
  • Smooth approximations with enough steepness recover the benefit, so the rule can be built with finite-slope comparators in analog hardware.
  • The improved performance comes with a simpler hardware requirement: reading signs is a one-bit local operation, whereas rescaling needs a global average over all spins.
  • The result unifies the treatment of external fields and higher-order couplings: one sign-based local-field construction mitigates imbalance at every interaction order.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper fixes one set of machine dynamics, so the natural next test is whether the sign rule still wins when momentum or chaotic amplitude control is added; that test would tell whether the ranking is a property of the local field or of the specific dynamics.
  • If the mechanism is order imbalance, the sign rule should show little or no advantage on problems with a single interaction order; the paper itself cites a study where it does worse in that setting, which provides a direct check.
  • A hardware-oriented prediction follows: the minimum steepness of a smooth sign approximation needed to match the exact rule is a measurable device specification, and machines with slower comparator transitions should see a smaller benefit.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper compares five ways of incorporating third-order (3-SAT) interactions into the local fields of a continuous-variable analog Ising machine. The dynamical model is the single-spin-rate equation dsi/dt = -si + tanh(alpha si + beta I_i), integrated by Euler-Maruyama with a linear annealing schedule for beta. On the first ten instances of six SATLIB uniform random 3-SAT families (N=20 to 250), the authors measure time-to-solution (TTS) and success rate (SR), with hyperparameters chosen by a per-instance grid search. They report that the 'spin sign' local field of Eq. (9), which replaces continuous spin amplitudes by their signs in the higher-order terms, yields lower TTS and higher SR than the baseline and three rescaling schemes, that its advantage grows with problem size, and that a smooth tanh(kappa s) approximation recovers the exact-sign performance as kappa grows. The conclusion is that spin-sign-based interactions are the most effective and hardware-compatible way to incorporate higher-order terms in analog Ising machines.

Significance. If the ranking is robust, the paper identifies a simple and practically attractive local-field prescription for higher-order terms, extending the authors' earlier result for external fields in quadratic models. The comparison is internally consistent: all methods share the same dynamics, the same hyperparameter grid, and the same public SATLIB instances, and the dt=0.01 integration step is cross-checked against dt=0.001 in the supplementary material. The smooth-approximation experiment (Fig. 3) is a useful step toward hardware realizability. However, the significance of the claim depends on two things that the manuscript does not fully establish: robustness across the dynamics actually used in analog Ising machines, and statistical reliability of the reported TTS/SR differences. The paper is honest about the fixed-dynamics scope in Section 4, but the abstract and conclusion are worded more broadly than the evidence supports.

major comments (3)
  1. [Abstract and Section 4, Eq. (4)] The headline claim that spin-sign interactions play a 'central role in enabling robust and scalable analog IM dynamics' is broader than the evidence presented. All results are obtained under one specific dynamical model: the gradient-based, momentum-free update of Eq. (4) with tanh saturation and a linear annealing schedule. The paper itself lists momentum [21,37] and chaotic amplitude control [43,46,47] as future work, and Kanao and Goto [37] report the opposite ranking for the spin-sign method under higher-order simulated-bifurcation dynamics. The authors' explanation for the discrepancy—that the prior study used a purely third-order model with no imbalances—is plausible but is not tested here. I agree with the authors' implicit scope limitation, but the abstract's general claim overreaches. Please either add experiments under at least one additional dynamics (e.g., with momentum or chaotic amplitude control) or explicitly restrict the conclusion to the tested dynamics class throughout the abstract and conclusion.
  2. [Section 5.2 and Figs. 1-2] The central empirical claim is reported without uncertainty quantification. TTS values are minima over the hyperparameter grid {alpha, v_beta, gamma} per instance and are estimated from 100 stochastic runs per hyperparameter setting, yet no confidence intervals, paired statistical tests, or sensitivity analyses are provided. Figures 1 and 2 report counts such as '54 of 60' and '59 of 60' as evidence of consistency, but with instance-level variability and with some methods returning TTS = infinity on several instances, these counts alone do not establish a statistically reliable ordering. Furthermore, Fig. S3 includes TTS = infinity values in boxplots, which can make distributions hard to interpret. Please report the variance of the optimized TTS (e.g., over repeated hyperparameter searches) and provide a paired significance test that handles censored TTS values, or clearly state that the conclusion is descriptive for this instance set.
  3. [Section 5.2, TTS definition] The time-to-solution metric is never defined explicitly. The text mentions that each TTS value corresponds to the optimal hyperparameter configuration and that 100 runs are used, but the standard TTS formula (which usually involves the time limit, the success probability, and a target success probability such as 0.99) is missing. Without this definition, the reported TTS values, the 'TTS = infinity' threshold, and the ordering in Figs. 1 and S2-S4 are not fully reproducible. Please include the exact formula and the rule for assigning TTS = infinity to unsolved configurations.
minor comments (5)
  1. [Figure 1 caption] The caption reads 'tmax = 10 4' with a missing superscript; it should be '10^4' to match the text in Section 5.2.
  2. [Section 2.1, Eqs. (6)-(8)] The notation <|s_m|> is not explicitly defined before first use; please state that the average is taken over all spin amplitudes in the machine at a given time.
  3. [Figure 2(b)] The panel labeled 'SR w.r.t. spin sign' should clarify whether it shows a difference or a ratio of success rates; the current y-axis label is ambiguous.
  4. [Section 6] The data availability statement says that additional data are available upon request, but no simulation code is released. For a numerical comparison whose central claim is a ranking of methods, releasing the code would materially aid reproducibility and is strongly encouraged.
  5. [Section 3.2, Eq. (10)] The phrase 'at κ = 10, it becomes the fastest method on the largest fraction of problems' is somewhat vague; please report the exact fraction or add a reference to a supplementary table.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the spin-sign advantage is an empirical benchmark result on external SATLIB instances, not a quantity fitted to the outcome or imported from self-citation.

full rationale

The paper's central claim is that the spin-sign local field of Eq. (9) outperforms four alternative local-field formulations on Uniform Random 3-SAT instances from SATLIB. This claim is established by direct numerical simulation of the dynamical equation (4) with each local field, measuring time-to-solution and success rate. No parameter is fitted to the performance metric and then renamed as a prediction; hyperparameters are optimized per instance for all methods symmetrically. The rescaling methods (Eqs. (6)-(8)) are derived from the baseline by construction, but their relative performance is not assumed and is instead measured. The spin-sign method itself is taken from prior work, but the paper tests it here rather than assuming its advantage; the citation to [21,37] is not used as evidence for the performance ranking. The authors' own prior work [35] is cited for context and for hardware-implementation remarks, but it is not load-bearing for the higher-order result. The only substantive limitation is the fixed no-momentum dynamics, which is an acknowledged scope condition, not a circular derivation. The numerical comparison is self-contained against external benchmarks, so there is no significant circularity.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central claim is empirical; it depends on the specific solver dynamics and benchmark protocol rather than on derived constants. No new physical entities are introduced; the spin sign local field is taken from prior work (Refs. 21, 37).

free parameters (2)
  • IM hyperparameters (alpha, v_beta, gamma) = alpha in [-10,1] integer steps; v_beta in {1e-4,1e-3,1e-2,1e-1,1}; gamma in {1e-4,1e-3,1e-2,1e-1}
    Chosen by hand and grid-searched per instance; TTS and SR are quoted at the best grid point, so the ranking depends on this set of values and on the fixed annealing schedule.
  • Time step dt = 0.01 (checked against 0.001 in supplementary)
    Discretization step for Euler-Maruyama; the supplementary comparison shows results are stable, so this is a mild setting rather than a fitted constant.
assumptions (4)
  • domain assumption Euler-Maruyama with dt = 0.01 faithfully integrates Eq. (4).
    The authors validate dt = 0.01 against 0.001 in Fig. S4; integration accuracy is assumed for all results.
  • standard math The PUBO embedding of Eqs. (12)-(14) maps satisfying assignments of a 3-SAT clause set to zero-energy ground states of the higher-order Ising Hamiltonian.
    An unsatisfied clause adds +1; the ground-state energy used as TTS target assumes this embedding is correct.
  • domain assumption The tanh nonlinearity in Eq. (4) is representative of analog Ising machine saturation behavior.
    The dynamics model is fixed across all methods; the ranking is conditional on this choice.
  • domain assumption Per-instance hyperparameter optimization gives a fair method comparison.
    TTS and SR are computed at the best grid point per instance; if the grid covers the methods differently, rankings could shift.

how reviews work

0 comments
Cite this review

Pith. "Pith review of How to Incorporate Higher-order Interactions in Analog Ising Machines." pith.science (2026). https://pith.science/paper/SX7PTJJO

@misc{pith2026250723621,
  author       = {Pith},
  title        = {Pith review of: How to Incorporate Higher-order Interactions in Analog Ising Machines},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SX7PTJJO}},
  note         = {Machine review of arXiv:2507.23621}
}
read the original abstract

Ising machines (IMs) are specialized devices designed to efficiently solve combinatorial optimization problems. Among such problems, Boolean Satisfiability (SAT) is particularly relevant in industrial applications. To solve SAT problems using IMs, it is crucial to incorporate higher-order interactions. However, in analog IMs, interactions of different orders scale unevenly with the continuous spin amplitudes, introducing imbalances that can significantly degrade performance. We present a numerical comparison of methods to mitigate these imbalances, evaluating time-to-solution and success rate on Uniform Random 3-SAT instances from the SATLIB benchmark set. Our results show that the most effective approach employs spin interactions that are proportional to the signs of spins, rather than their continuous amplitudes. This generalizes our previous work, which showed that such interactions best mitigate imbalances induced by external fields in quadratic analog IMs. In this work, its advantage becomes substantially more pronounced, as it naturally mitigates imbalances across all interaction orders. We further demonstrate that smooth approximations of this method make it compatible with analog hardware. Our findings underscore the central role of spin-sign-based interactions in enabling robust and scalable analog IM dynamics.

Figures

Figures reproduced from arXiv: 2507.23621 by the authors.

Figure 1
Figure 1. Comparison of TTS between the methods to incorporate higher-order interactions of Section 2.1 for Uniform Random-3-SAT problems. Dots in the grey area on the right denote COPs that could be solved by the spin sign method within the allocated compute time of tmax = 104 , but not by the method on the x-axis (TTS = ∞, SR = 0). The spin sign method generally requires less time to solve problems, and solves more problems… view at source ↗
Figure 2
Figure 2. Comparison of success rates across the different methods to incorporate higher-order interactions of Section 2.1 on Uniform Random-3-SAT problems. Panel (a) shows absolute SRs, while panel (b) displays SRs relative to the spin sign method. Each boxplot represents a distribution over 10 instances of the same size for a given method, where dots indicate outliers. A detailed breakdown for individual problem instances i… view at source ↗
Figure 3
Figure 3. Fraction of benchmark problems for which each method is fastest, as a function of κ in the smooth tanh(κsi) approximation of the spin sign method. As the performance of the approximate method improves with increasing κ, the relative rank￾ing across all methods shifts, eventually reaching the limit of the exact spin sign method. 6 [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (1 more)
Figure 1
Figure 1. Figure 1: We compare the various methods to incorporate higher-order interactions, as described in Section 2.1. [PITH_FULL_IMAGE:figures/full_fig_p015_1.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. A More Convex Ising Formulation of Max-3-Cut Using Higher-Order Spin Interactions

    cond-mat.stat-mech 2025-08 conditional novelty 7.0 of 10

    A four-spin Ising formulation for Max-3-Cut removes the energy barriers of one-hot encoding and is about 2.75x faster than a rescaled quadratic baseline on an analog Ising simulator.

  2. Photonic Ising machines toward and beyond a million spins

    physics.optics 2026-07 conditional novelty 4.0 of 10

    Million-spin photonic Ising machines are argued to be within reach via chiplet, free-space, and all-optical spatiotemporal architectures, but only with major engineering advances.

Reference graph

Works this paper leans on

50 extracted references · 47 canonical work pages · cited by 2 Pith papers

  1. [37]

    Simulated bifur- cation for higher-order cost functions

    Taro Kanao and Hayato Goto. Simulated bifur- cation for higher-order cost functions. Applied Physics Express, 16(1):014501, 2022

  2. [1]

    Using a sat-solver to schedule sports leagues

    Andrei Horbach, Thomas Bartsch, and Dirk Briskorn. Using a sat-solver to schedule sports leagues. Journal of Scheduling , 15:117–125, 2012

  3. [2]

    Randomization in backtrack search: Exploiting heavy-tailed profiles for solv- ing hard scheduling problems

    Carla P Gomes, Bart Selman, Ken McAloon, and Carol Tretkoff. Randomization in backtrack search: Exploiting heavy-tailed profiles for solv- ing hard scheduling problems. In AIPS, pages 208–213, 1998

  4. [3]

    Pushing the envelope: Planning, propositional logic, and stochastic search

    Henry Kautz and Bart Selman. Pushing the envelope: Planning, propositional logic, and stochastic search. In Proceedings of the national conference on artificial intelligence, pages 1194– 1201, 1996

  5. [4]

    Artificial in- telligence: a modern approach

    Stuart J Russell and Peter Norvig. Artificial in- telligence: a modern approach . pearson, 2016

  6. [5]

    Symbolic model checking: 1020 states and be- yond

    Jerry R Burch, Edmund M Clarke, Kenneth L McMillan, David L Dill, and Lain-Jinn Hwang. Symbolic model checking: 1020 states and be- yond. Information and computation , 98(2):142– 170, 1992

  7. [6]

    Symbolic model checking using sat procedures instead of bdds

    Armin Biere, Alessandro Cimatti, Edmund M Clarke, Masahiro Fujita, and Yunshan Zhu. Symbolic model checking using sat procedures instead of bdds. In Proceedings of the 36th annual ACM/IEEE Design Automation Confer- ence, pages 317–320, 1999

  8. [7]

    Effec- tive use of boolean satisfiability procedures in the formal verification of superscalar and vliw

    Miroslav N Velev and Randal E Bryant. Effec- tive use of boolean satisfiability procedures in the formal verification of superscalar and vliw. In Proceedings of the 38th annual design automa- tion conference, pages 226–231, 2001

Show all 50 references
  1. [8]

    Explorations of sequential atpg using boolean satisfiability

    Haluk Konuk Tracy Larrabee et al. Explorations of sequential atpg using boolean satisfiability. In VLSI Test Symposium 1993 , pages 85–90. Cite- seer, 1993

  2. [9]

    Combinational test generation using satisfiability

    Paul Stephan, Robert K Brayton, and Al- berto L Sangiovanni-Vincentelli. Combinational test generation using satisfiability. IEEE Trans- actions on Computer-Aided Design of Integrated Circuits and Systems , 15(9):1167–1176, 2002

  3. [10]

    Solving olympiad geome- try without human demonstrations

    Trieu H Trinh, Yuhuai Wu, Quoc V Le, He He, and Thang Luong. Solving olympiad geome- try without human demonstrations. Nature, 625 (7995):476–482, 2024

  4. [11]

    Fine-tuning language models us- ing formal methods feedback: A use case in autonomous systems

    Yunhao Yang, Neel P Bhatt, Tyler Ingebrand, William Ward, Steven Carr, Atlas Wang, and Ufuk Topcu. Fine-tuning language models us- ing formal methods feedback: A use case in autonomous systems. Proceedings of Machine Learning and Systems , 6:339–350, 2024

  5. [12]

    SA Cook. Proc. 3rd annu. acm symp. on theory of computing, 1971

  6. [13]

    The cook-levin theorem

    Frank J Balbach. The cook-levin theorem. Archive of Formal Proofs, 5:428, 2023

  7. [14]

    Ising machines as hardware solvers of combinatorial optimization problems

    Naeimeh Mohseni, Peter L McMahon, and Tim Byrnes. Ising machines as hardware solvers of combinatorial optimization problems. Nature Reviews Physics, 4(6):363–379, 2022

  8. [15]

    Probabilistic computing with p-bits

    Jan Kaiser and Supriyo Datta. Probabilistic computing with p-bits. Applied Physics Let- ters, 119, 2021. doi: https://doi.org/10.1063/ 5.0067927

  9. [16]

    Large- scale photonic ising machine by spatial light modulation

    D Pierangeli, G Marcucci, and C Conti. Large- scale photonic ising machine by spatial light modulation. Physical review letters , 122(21): 213902, 2019. 9

  10. [17]

    Power-efficient combinatorial optimization using intrinsic noise in memristor hopfield neural net- works

    Fuxi Cai, Suhas Kumar, Thomas Van Vaeren- bergh, Xia Sheng, Rui Liu, Can Li, Zhan Liu, Martin Foltin, Shimeng Yu, Qiangfei Xia, et al. Power-efficient combinatorial optimization using intrinsic noise in memristor hopfield neural net- works. Nature Electronics, 3(7):409–418, 2020

  11. [18]

    Quantum an- nealing with manufactured spins

    Mark W Johnson, Mohammad HS Amin, Suzanne Gildert, Trevor Lanting, Firas Hamze, Neil Dickson, Richard Harris, Andrew J Berkley, Jan Johansson, Paul Bunyk, et al. Quantum an- nealing with manufactured spins. Nature, 473 (7346):194–198, 2011

  12. [19]

    100,000-spin coherent ising machine

    Toshimori Honjo, Tomohiro Sonobe, Kensuke Inaba, Takahiro Inagaki, Takuya Ikuta, Ya- suhiro Yamada, Takushi Kazama, Koji En- butsu, Takeshi Umeki, Ryoichi Kasahara, et al. 100,000-spin coherent ising machine. Science ad- vances, 7(40):eabh0952, 2021

  13. [20]

    Destabiliza- tion of local minima in analog spin systems by correction of amplitude heterogeneity

    Timoth´ ee Leleu, Yoshihisa Yamamoto, Peter L McMahon, and Kazuyuki Aihara. Destabiliza- tion of local minima in analog spin systems by correction of amplitude heterogeneity. Physical review letters, 122(4):040607, 2019

  14. [21]

    High-performance combinatorial opti- mization based on classical mechanics

    Hayato Goto, Kotaro Endo, Masaru Suzuki, Yoshisato Sakai, Taro Kanao, Yohei Hamakawa, Ryo Hidaka, Masaya Yamasaki, and Kosuke Tat- sumura. High-performance combinatorial opti- mization based on classical mechanics. Science Advances, 7(6):eabe7953, 2021

  15. [22]

    A poor man’s coherent ising machine based on opto-electronic feedback sys- tems for solving optimization problems

    Fabian B¨ ohm, Guy Verschaffelt, and Guy Van der Sande. A poor man’s coherent ising machine based on opto-electronic feedback sys- tems for solving optimization problems. Nature communications, 10(1):3538, 2019

  16. [23]

    L. Q. English, A. V. Zampetaki, K. P. Kalinin, N. G. Berloff, and P. G. Kevrekidis. An ising machine based on networks of subharmonic elec- trical resonators. Communications Physics , 5, 2022

  17. [24]

    Efficient combinatorial optimization by quantum-inspired parallel annealing in analogue memristor crossbar

    Mingrui Jiang, Keyi Shan, Chengping He, and Can Li. Efficient combinatorial optimization by quantum-inspired parallel annealing in analogue memristor crossbar. Nature communications, 14 (1):5927, 2023

  18. [25]

    Global optimization of spin hamiltonians with gain- dissipative systems

    Kirill P Kalinin and Natalia G Berloff. Global optimization of spin hamiltonians with gain- dissipative systems. Scientific reports , 8(1): 17791, 2018

  19. [26]

    Soft constraints

    Pedro Meseguer, Francesca Rossi, and Thomas Schiex. Soft constraints. In Foundations of Arti- ficial Intelligence , volume 2, pages 281–328. El- sevier, 2006

  20. [27]

    Energy land- scapes of combinatorial optimization in ising ma- chines

    Dmitrii Dobrynin, Adrien Renaudineau, Mo- hammad Hizzani, Dmitri Strukov, Masoud Mohseni, and John Paul Strachan. Energy land- scapes of combinatorial optimization in ising ma- chines. Physical Review E , 110(4):045308, 2024

  21. [28]

    Efficient optimization with higher-order ising machines

    Connor Bybee, Denis Kleyko, Dmitri E Nikonov, Amir Khosrowshahi, Bruno A Olshausen, and Friedrich T Sommer. Efficient optimization with higher-order ising machines. Nature Communi- cations, 14(1):6033, 2023

  22. [29]

    Compu- tational overhead of locality reduction in binary optimization problems

    Elisabetta Valiante, Maritza Hernandez, Amin Barzegar, and Helmut G Katzgraber. Compu- tational overhead of locality reduction in binary optimization problems. Computer Physics Com- munications, 269:108102, 2021

  23. [30]

    Memristor-based hardware and al- gorithms for higher-order hopfield optimization solver outperforming quadratic ising machines

    Mohammad Hizzani, Arne Heittmann, George Hutchinson, Dmitrii Dobrynin, Thomas Van Vaerenbergh, Tinish Bhattacharya, Adrien Renaudineau, Dmitri Strukov, and John Paul Strachan. Memristor-based hardware and al- gorithms for higher-order hopfield optimization solver outperforming...

  24. [31]

    10 Solving boolean satisfiability problems with re- sistive content addressable memories

    Giacomo Pedretti, Fabian B¨ ohm, Tinish Bhat- tacharya, Arne Heittmann, Xiangyi Zhang, Mo- hammad Hizzani, George Hutchinson, Dongseok Kwon, John Moon, Elisabetta Valiante, et al. 10 Solving boolean satisfiability problems with re- sistive content addressable memories. npj Un-...

  25. [32]

    Boltzmann sampling by degenerate optical parametric oscillator network for structure-based virtual screening

    Hiromasa Sakaguchi, Koji Ogata, Tetsu Iso- mura, Shoko Utsunomiya, Yoshihisa Yamamoto, and Kazuyuki Aihara. Boltzmann sampling by degenerate optical parametric oscillator network for structure-based virtual screening. Entropy, 18(10):365, 2016

  26. [33]

    Ising ma- chines’ dynamics and regularization for near- optimal mimo detection

    Abhishek Kumar Singh, Kyle Jamieson, Peter L McMahon, and Davide Venturelli. Ising ma- chines’ dynamics and regularization for near- optimal mimo detection. IEEE Transactions on Wireless Communications, 21(12):11080–11094, 2022

  27. [34]

    Mean-field coher- ent ising machines with artificial zeeman terms

    Sudeera Hasaranga Gunathilaka Mastiyage Don, Yoshitaka Inui, Satoshi Kako, Yoshihisa Ya- mamoto, and Toru Aonishi. Mean-field coher- ent ising machines with artificial zeeman terms. Journal of Applied Physics , 134(23), 2023

  28. [35]

    How to incorporate external fields in analog ising machines

    Robbe De Prins, Jacob Lamers, Peter Bienst- man, Guy Van der Sande, Guy Verschaffelt, and Thomas Van Vaerenbergh. How to incorporate external fields in analog ising machines. arXiv preprint arXiv:2505.08796, 2025

  29. [36]

    Polynomial uncon- strained binary optimisation inspired by opti- cal simulation

    Dmitry A Chermoshentsev, Aleksei O Malyshev, Mert Esencan, Egor S Tiunov, Douglas Men- doza, Al´ an Aspuru-Guzik, Aleksey K Fedorov, and Alexander I Lvovsky. Polynomial uncon- strained binary optimisation inspired by opti- cal simulation. arXiv preprint arXiv:2106.13167, 2021

  30. [38]

    Hoos and T

    H. Hoos and T. Stutzle. Satlib: An online re- source for research on sat. In H. van Maaren I. P. Gent and T. Walsh, editors, SAT2000, pages 283–292. IOS Press, 2000

  31. [39]

    Geometric landscape annealing as an optimization principle underlying the coherent ising machine

    Atsushi Yamamura, Hideo Mabuchi, and Surya Ganguli. Geometric landscape annealing as an optimization principle underlying the coherent ising machine. Phys. Rev. X , 14:031054, 2024. doi: 10.1103/PhysRevX.14.031054

  32. [40]

    Using continuation methods to analyse the difficulty of problems solved by ising machines

    Jacob Lamers, Guy Verschaffelt, and Guy Van der Sande. Using continuation methods to analyse the difficulty of problems solved by ising machines. Communications Physics, 7, 2024

  33. [41]

    Order-of- magnitude differences in computational perfor- mance of analog ising machines induced by the choice of nonlinearity

    Fabian B¨ ohm, Thomas Van Vaerenbergh, Guy Verschaffelt, and Guy Van der Sande. Order-of- magnitude differences in computational perfor- mance of analog ising machines induced by the choice of nonlinearity. Communications Physics, 4(1):149, 2021

  34. [42]

    Using a spatial light modu- lator to achieve massive-parallelization in ana- log ising machines

    Toon Sevenants. Using a spatial light modu- lator to achieve massive-parallelization in ana- log ising machines. Conference talk at the CLEO/Europe-EQEC, 2025

  35. [43]

    Coherent sat solvers: a tutorial

    Sam Reifenstein, Timothee Leleu, Timothy McKenna, Marc Jankowski, Myoung-Gyun Suh, Edwin Ng, Farad Khoyratee, Zoltan Toroczkai, and Yoshihisa Yamamoto. Coherent sat solvers: a tutorial. Advances in Optics and Photonics , 15(2):385–441, 2023

  36. [44]

    A coherent ising machine for 2000-node opti- mization problems

    Takahiro Inagaki, Yoshitaka Haribara, Koji Igarashi, Tomohiro Sonobe, Shuhei Tamate, Toshimori Honjo, Alireza Marandi, Peter L McMahon, Takeshi Umeki, Koji Enbutsu, et al. A coherent ising machine for 2000-node opti- mization problems. Science, 354(6312):603–606, 2016

  37. [45]

    Realizing the classical xy hamilto- nian in polariton simulators

    Natalia G Berloff, Matteo Silva, Kirill Kalinin, Alexis Askitopoulos, Julian D T¨ opfer, Pasquale Cilibrizzi, Wolfgang Langbein, and Pavlos G Lagoudakis. Realizing the classical xy hamilto- nian in polariton simulators. Nature materials , 16(11):1120–1126, 2017

  38. [46]

    Destabiliza- tion of local minima in analog spin systems by correction of amplitude heterogeneity

    Timoth´ ee Leleu, Yoshihisa Yamamoto, Peter L McMahon, and Kazuyuki Aihara. Destabiliza- tion of local minima in analog spin systems by correction of amplitude heterogeneity. Physical review letters, 122(4):040607, 2019. 11

  39. [47]

    Scaling advantage of chaotic amplitude control for high-performance combi- natorial optimization

    Timoth´ ee Leleu, Farad Khoyratee, Timoth´ ee Levi, Ryan Hamerly, Takashi Kohno, and Kazuyuki Aihara. Scaling advantage of chaotic amplitude control for high-performance combi- natorial optimization. Communications Physics, 4(1):266, 2021

  40. [48]

    Designing ising machines with higher order spin interactions and their application in solving combinatorial optimization

    Mohammad Khairul Bashar and Nikhil Shukla. Designing ising machines with higher order spin interactions and their application in solving combinatorial optimization. Scientific Reports, 13(1):9558, 2023

  41. [49]

    On the empiri- cal time complexity of random 3-sat at the phase transition

    Zongxu Mu and Holger H Hoos. On the empiri- cal time complexity of random 3-sat at the phase transition. In IJCAI, pages 367–373, 2015

  42. [50]

    Analysing the empirical time com- plexity of high-performance algorithms for SAT and TSP

    Zongxu Mu. Analysing the empirical time com- plexity of high-performance algorithms for SAT and TSP . PhD thesis, University of British Columbia, 2015. 12 Supplementary Material S.1 Success rates and time- to-solution for individual problem instances Fig. S1 shows the success ...

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.