Astroid Spinodal Boundary in Phase-Based Ising Machines

Malihe Farasat; Nikhil Shukla

REVIEW 2 major objections 5 minor 19 references

Astroid Spinodal Boundary in Phase-Based Ising Machines

T0 review · 2 major / 5 minor · reviewed 2026-08-01 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The conditional energy landscape of a single oscillator in phase-based Ising machines is bistable inside an astroid-shaped boundary, not merely at a critical SHI strength.

desk verdict A clean, honestly framed derivation of the astroid spinodal for the conditional single-oscillator landscape; the gap to full network dynamics is real but not fatal because the paper never claims more than it proves. read the letter →

arxiv 2607.26213 v1 pith:6TF5C2GJ submitted 2026-07-28 physics.comp-ph

Malihe Farasat , Nikhil Shukla This is my paper · ORCID

classification physics.comp-ph

keywords OscillatorIsingmachinesDynamicalspinodalboundaryastroidbistabilitysecond-harmonicinjectionStoner–Wohlfarthenergybarrier

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

The paper claims that in oscillator Ising machines (OIMs) and dynamical Ising machines (DIMs), the competition between second-harmonic injection (SHI) and the instantaneous local network field reshapes each oscillator's conditional energy landscape, and that this landscape undergoes a monostable-to-bistable transition when the normalized local field crosses an astroid spinodal boundary. The paper derives the astroid boundary from a fold condition and shows the barrier vanishes with exponent 3/2 at generic smooth points and exponent 2 at the longitudinal cusp. Because OIMs and DIMs share the same geometry, and the first-harmonic landscape is mathematically the Stoner–Wohlfarth energy, the result connects phase-based Ising machine design to a classical magnet physics framework. The matter-of-course consequence is that whether a node can get trapped in a wrong phase state is not set by coupling strength alone but by the full orientation of the local field relative to the astroid.

What carries the argument

The load-bearing object is the astroid spinodal boundary, obtained by simultaneously imposing stationarity and vanishing curvature on the one-dimensional conditional landscape. Parametrically, the boundary is h_x,i = -2K_s cos^3 θ_sp and h_y,i = 2K_s sin^3 θ_sp, which eliminate to the astroid equation. This object organizes the entire argument: it fixes the SHI threshold K_s,tr,i = (1/2)(|h_x|^{2/3}+|h_y|^{2/3})^{3/2}, determines barrier scaling through a fold expansion, and connects the oscillator problem to the Stoner–Wohlfarth switching astroid.

What would settle it

The decisive check is a numerical integration of a small coupled OIM or DIM network where the local field is allowed to evolve dynamically, compared against the frozen-field astroid prediction: if nodes spend time inside the frozen astroid but do not exhibit two stable wells, or vice versa, the quasi-static boundary is violated. Alternatively, escape-time measurements near a smooth crossing versus near the longitudinal cusp could test the predicted 3/2 versus 2 barrier scaling.

Watch

Extended reading notes

Core claim

The central discovery is that local bistability in phase-based Ising machines is governed by an astroid spinodal boundary in the normalized local-field plane. For a fixed snapshot of neighboring phases, the conditional energy E_i(θ_i) = -h_x,i cos θ_i - h_y,i sin θ_i - (K_s/2) cos 2θ_i has two stable wells exactly when |h_x,i/(2K_s)|^{2/3} + |h_y,i/(2K_s)|^{2/3} < 1. Crossing this boundary creates or destroys a metastable well, with barrier height scaling as K_s μ_i^{3/2} at smooth points and K_s μ_i^2 at the longitudinal cusp, where μ_i is the spinodal margin. The paper also shows OIMs and DIMs differ only by reflection of the transverse field and that the same energy is equivalent (up to a

Load-bearing premise

The paper assumes the local network field (h_x,i, h_y,i) is frozen when computing the spinodal boundary and barrier, whereas in a real coupled network this field evolves together with the oscillator's phase; if it fluctuates quickly, the quasi-static boundary may not describe actual bistability.

Editorial extensions

If this is right

The SHI threshold for local bistability is not a single coupling value but a field-orientation-dependent threshold; equal-magnitude fields can require SHI strengths differing by a factor of two.
Since OIMs and DIMs share the astroid geometry, any design rule for avoiding metastable traps applies to both architectures with the transverse field sign reversed.
The Stoner–Wohlfarth equivalence lets phase-based Ising machine bistability be analyzed with known results from single-domain magnet switching, including barrier formulas and switching probabilities.
The barrier scaling (3/2 at smooth points, 2 at cusp) determines how quickly metastable wells become effective traps as the SHI strength is increased beyond threshold.
In a network, nodes repeatedly cross the astroid during dynamics, so local monostable/bistable transitions are a normal part of computation, not just an endpoint effect.

Reading between the lines

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

Because the local field is treated as frozen only for the spinodal analysis, the astroid is a quasi-static boundary; in fast networks the effective boundary may be smeared or shifted by field fluctuations, a testable extension being comparison with simulations that include finite relaxation times.
The phasor-sum expression for the local field suggests that graph structure enters through interference, not just degree; wiring loops or motifs that promote partial phase cancellation could systematically delay nodes from entering the bistable region, an effect the paper leaves implicit.
The same first-harmonic landscape appears in other phase-encoded computing schemes, so the astroid criterion could be ported to any system described by a cosine double-well plus a first-harmonic tilt, including some optical or mechanical Ising machines.
The cusp scaling ΔE ∝ μ^2 is slower than the generic 3/2 scaling, implying that longitudinal crossing events create barriers that grow quadratically; one could test this by intentionally biasing nodes along the axis and measuring escape times.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

Quick take: the astroid boundary for the conditional one-oscillator energy is derived correctly, the SHI threshold formula is new and useful, and the Stoner–Wohlfarth mapping is a genuinely nice observation. The paper is careful to say all results are for a fixed snapshot of neighboring phases, so the main soft spot is the unvalidated leap from that frozen-field picture to what a coupled network actually does.

The new material: parametric spinodal (Eq. 19), the astroid boundary (Eq. 20), the generalized SHI threshold (Eq. 22), and the barrier scaling exponents 3/2 and 2 for generic points and the longitudinal cusp. I verified the key algebra myself; it checks out. The paper also cleanly separates longitudinal and transverse fields, shows OIMs and DIMs differ only by transverse-field reversal, and identifies the exact equivalence with the Stoner–Wohlfarth energy. That equivalence is not just a metaphor—the normalized energies are identical up to an additive constant. Good, solid work.

The soft spots are proportionate. The frozen-field assumption is the big one. The stress-test objection about the full N-oscillator Jacobian is mathematically correct, but it doesn't fully land because the paper never claims the astroid is a bifurcation set of the coupled system. The claims are explicitly about the conditional landscape. The legitimate issue is that the practical significance for choosing SHI strength depends on whether the quasi-static picture is a good approximation in actual dynamics, and the paper doesn't test that. Fig. 3 only plots field trajectories against the astroid; it doesn't show that a crossing coincides with a saddle-node event or that the predicted barrier scaling is observed. The graph-structure discussion is heuristic. There's also no code or data beyond the illustrative figure—"available upon reasonable request" is the standard placeholder.

Who this is for: people working on oscillator Ising machines or dynamical Ising machines who want design intuition about SHI strength, and anyone interested in exact analogies between phase-based spin systems and micromagnetic energy landscapes. It's not a breakthrough, but it's a well-made, self-contained contribution.

Verdict: yes, this deserves a serious referee. The math is sound and the SW hook is valuable. A revision should either add a small dynamical validation (e.g., a few-node simulation showing that crossing the astroid correlates with the appearance or disappearance of a metastable state in the full flow) or explicitly temper the framing from "drives a transition in the machine" to "describes the conditional landscape." I'd engage with it.

Referee Report

2 major / 5 minor

Summary. The paper analyzes the instantaneous conditional energy E_i(θ_i) = −h_x,i cosθ_i − h_y,i sinθ_i − (K_s/2) cos(2θ_i) of a single oscillator in OIM/DIM networks. It derives the astroid spinodal boundary (Eq. 20), the SHI threshold (Eq. 22), and the barrier scalings ΔE ∝ μ^{3/2} at generic smooth points and ΔE ∝ μ^2 at the longitudinal cusp (Eqs. 35, 41). It also proves an exact correspondence with the Stoner–Wohlfarth energy and shows that OIM and DIM conditional landscapes differ only by a reflection of the transverse field. The derivation is self-contained and algebraically correct within the stated frozen-field framework.

Significance. The result is a clean, parameter-free characterization of local bistability in the conditional landscape of phase-based Ising machines. The Stoner–Wohlfarth correspondence is exact and provides a useful unifying perspective. The barrier scalings are derived from a normal-form expansion rather than fitted, and the astroid boundary is an explicit algebraic result. The main significance is as a local, quasi-static analysis; its direct applicability to the full coupled network dynamics is not established and requires scoping or numerical support.

major comments (2)

[Sec. II, Eq. (2), Eq. (20), Fig. 3] The astroid is derived from the 1D conditional energy with h_x,i, h_y,i frozen. This is a spinodal of the conditional landscape, but the paper's language ('locally stable', 'trapped', 'bistability in phase-based Ising machines') implies statements about the coupled N-oscillator system. In the full dynamics the local field co-evolves; a zero of ∂²E_i/∂θ_i² at a conditional stationary point is not a zero eigenvalue of the full Jacobian because off-diagonal couplings shift stability. Fig. 3 only plots field trajectories against the astroid; it does not show that crossings coincide with saddle-node events of the full system or that escape rates follow Eqs. (35)/(41). Please either justify a timescale separation or explicitly scope all claims to the frozen-field conditional landscape.
[Sec. II.B, Eq. (21)] The paper states that inside the astroid the landscape is bistable and outside monostable. This is plausible—extrema can only appear/disappear at degenerate critical points on the astroid, and h=0 has two minima—but no proof is given. A short argument, or a reference to the Stoner–Wohlfarth analysis, is needed to establish that the number of minima is two for all ℓ<1 and one for ℓ>1. This is load-bearing because the central claim of the paper is that Eq. (20) is the boundary separating monostable from bistable regimes.

minor comments (5)

[Eq. (39)] The notation E_i(q)−E_i(0) with q=θ_i−π conflicts with Eq. (5), where E_i(0) means E_i(θ_i=0). Here E_i(0) is the value at q=0 (i.e., θ_i=π). Please relabel, e.g., E_i(π+q)−E_i(π).
[Sec. II.A, after Eq. (17)] The statement that increasing |h_x,i| 'eventually restores a single-well landscape' is true, but the threshold |h_x,i|=2K_s is not stated. Adding it would make the on-axis discussion more precise.
[Fig. 3] The caption says 'randomly generated graph with N nodes and edge density p' while the text gives N=10, p=0.5. Please state the graph construction in the caption and describe how the crossing points ('diamonds') are detected.
[Introduction / abstract] The term 'first-harmonic conditional landscape' in the abstract is not defined. Clarify that it refers to the energy containing cosθ and sinθ terms plus the SHI cos2θ term.
[Acknowledgments] Minor typographical issue: 'NSF grant #No. 2328961' should read 'NSF grant No. 2328961'.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular dependency; the astroid and barrier scalings are algebraic consequences of the stated energy model.

full rationale

The derivation is self-contained. Eq. (2) defines the instantaneous conditional energy from the network coupling and SHI; Eqs. (3)-(4) are an exact rearrangement. The spinodal condition (18)-(19) imposes stationarity and vanishing curvature of this one-dimensional energy, and eliminating θ_sp gives the astroid (20). No parameter is fitted to produce the astroid: K, K_s, and W_ij enter only as model inputs, and the boundary is the analytic fold set of the stated potential. The threshold (22) is just the astroid equation solved for K_s. The barrier scalings (35) and (41) follow from a generic fold expansion in the spinodal margin μ_i, defined via (30)-(31); the μ^(3/2) and μ^2 exponents are singularity-theory consequences of the same energy, not imported from data or from prior results. The Stoner–Wohlfarth correspondence in Sec. III is an equivalence noted after the astroid is derived, not used to obtain it; references [17-19] are independent classical results. The only self-citations [10,15] provide background and the OIM/DIM model definitions, not the astroid, barrier, or SW mapping, so they are not load-bearing. The conditional/frozen-neighbor-field scope is a modeling limitation rather than a circular identification: the paper does not claim to fit h fields from the barrier, and the barrier formula is derivative of the same energy, not an input. Thus no circular step can be exhibited.

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

No free parameters are fitted; model inputs (K, K_s, W_ij) are physical constants of the setup. The central derivation rests on the energy model (Eq. 2), the gradient-flow relaxation assumption, and the frozen-field approximation. No new entities (particles, forces, dimensions) are introduced.

assumptions (4)

domain assumption The conditional energy of oscillator i is given by Eq. (2): E_i = −K Σ_j W_ij cos(θ_i + σ θ_j) − (K_s/2) cos(2θ_i).
This is the standard model for OIM (σ=−1) and DIM (σ=+1), taken from refs. [14] and [15]. The entire spinodal derivation operates within this model; if the actual oscillator dynamics are not governed by this energy, the results would not apply.
domain assumption The oscillator phase relaxes (overdamped) on the conditional energy landscape, so minima/maxima of E_i determine stability.
Implicit throughout the paper; the analysis of monostability/bistability via stationary points and curvature assumes that the phase dynamics are effectively a gradient flow on E_i. Not explicitly stated.
domain assumption The local fields h_x,i and h_y,i are frozen when analyzing the spinodal and barrier (quasi-static approximation).
Explicitly stated in Sec. II: 'For a fixed snapshot of the neighboring phases...' The paper does not discuss the validity of this approximation for a coupled network where the fields evolve; this is a load-bearing assumption for the instantaneous-landscape picture.
standard math Standard catastrophe-theory local expansions (fold and cusp) apply at spinodal points, with the required non-degeneracy conditions.
Used in Sec. II-C to derive the μ^{3/2} and μ^2 barrier scalings. The paper asserts ∂^3E ≠ 0 at generic smooth points and the corresponding cusp condition but does not exhaustively verify these for all directions.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Astroid Spinodal Boundary in Phase-Based Ising Machines." pith.science (2026). https://pith.science/paper/6TF5C2GJ

@misc{pith2026260726213,
  author       = {Pith},
  title        = {Pith review of: Astroid Spinodal Boundary in Phase-Based Ising Machines},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6TF5C2GJ}},
  note         = {Machine review of arXiv:2607.26213}
}

abstract

Oscillator Ising machines (OIMs) and dynamical Ising machines (DIMs) encode binary spins in phase states stabilized by second-harmonic injection (SHI). In a coupled network, the competition between SHI and the instantaneous local network field reshapes each oscillator's conditional energy landscape. We show that this competition drives a transition between monostable and bistable regimes through an astroid spinodal boundary. Near this boundary, the barrier scales as $\Delta E_i\propto\mu_i^{3/2}$ at generic smooth points and as $\Delta E_i\propto\mu_i^{2}$ at the longitudinal cusp. OIMs and DIMs obey the same spinodal geometry, with their conditional landscapes related by a reversal of the transverse field. Finally, the first-harmonic conditional landscape is mathematically equivalent, up to an additive constant, to the Stoner--Wohlfarth energy of a uniaxial magnetic particle.

Figures

Figures reproduced from arXiv: 2607.26213 by the authors.

**Figure 2.** (a) Schematic of the spinodal astroid in the normalized local-field plane [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** Trajectories of the normalized local fields [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

19 extracted references

[1]

Mohseni, P

N. Mohseni, P. L. McMahon, and T. Byrnes, Ising ma- chines as hardware solvers of combinatorial optimization problems, Nature Reviews Physics4, 363 (2022)

2022
[2]

Zhang, Q

T. Zhang, Q. Tao, B. Liu, A. Grimaldi, E. Raimondo, M. Jim´ enez, M. J. Avedillo, J. Nunez, B. Linares- Barranco, T. Serrano-Gotarredona, G. Finocchio, and J. Han, A review of Ising machines implemented in con- ventional and emerging technologies, IEEE Transactions on Nanotechnology23, 704 (2024)

2024
[3]

Lucas, Ising formulations of many NP problems, Fron- tiers in Physics2, 74887 (2014)

A. Lucas, Ising formulations of many NP problems, Fron- tiers in Physics2, 74887 (2014)

2014
[4]

Chang, C.-F

Y.-J. Chang, C.-F. Nien, K.-P. Huang, Y.-T. Zhang, C.-H. Cho, and C.-R. Chang, Quantum computing for optimization with Ising machine, IEEE Nanotechnology Magazine18, 15 (2024)

2024
[5]

P. L. McMahon, A. Marandi, Y. Haribara, R. Hamerly, C. Langrock, S. Tamate, T. Inagaki, H. Takesue, S. Ut- sunomiya, K. Aihara,et al., A fully programmable 100- spin coherent Ising machine with all-to-all connections, Science354, 614 (2016)

2016
[6]

Ezawa, E

M. Ezawa, E. Lebrasseur, and Y. Mita, Ising machine based on bistable microelectromechanical systems, Jour- nal of the Physical Society of Japan91, 114601 (2022)

2022
[7]

Grimaldi, L

A. Grimaldi, L. Mazza, E. Raimondo, P. Tullo, D. Ro- drigues, K. Y. Camsari, V. Crupi, M. Carpentieri, V. Puliafito, and G. Finocchio, Evaluating spintronics- compatible implementations of Ising machines, Physical Review Applied20, 024005 (2023)

2023
[8]

J. Chou, S. Bramhavar, S. Ghosh, and W. Herzog, Analog coupled oscillator based weighted Ising machine, Scien- tific Reports9, 14786 (2019)

2019

Show all 19 references

[9]

Mallick, Z

A. Mallick, Z. Zhao, M. K. Bashar, S. Alam, M. M. Islam, Y. Xiao, Y. Xu, A. Aziz, V. Narayanan, K. Ni,et al., CMOS-compatible Ising machines built using bistable latches coupled through ferroelectric transistor arrays, Scientific Reports13, 1515 (2023)

2023
[10]

M. K. Bashar, A. Mallick, and N. Shukla, Experimental investigation of the dynamics of coupled oscillators as Ising machines, IEEE Access9, 148184 (2021)

2021
[11]

W. Moy, I. Ahmed, P.-w. Chiu, J. Moy, S. S. Sapatnekar, and C. H. Kim, A 1,968-node coupled ring oscillator cir- cuit for combinatorial optimization problem solving, Na- ture Electronics5, 310 (2022)

2022
[12]

Afoakwa, Y

R. Afoakwa, Y. Zhang, U. K. R. Vengalam, Z. Ignja- tovic, and M. Huang, BRIM: Bistable resistively-coupled Ising machine, in2021 IEEE International Symposium on High-Performance Computer Architecture (HPCA) (IEEE, 2021) pp. 749–760

2021
[13]

Casilli, T

N. Casilli, T. Kaisar, L. Colombo, S. Ghosh, P. X.- L. Feng, and C. Cassella, Parametric frequency divider based ising machines, Phys. Rev. Lett.132, 147301 (2024)

2024
[14]

Wang and J

T. Wang and J. Roychowdhury, OIM: Oscillator-based Ising machines for solving combinatorial optimisation problems, inInternational Conference on Unconventional Computation and Natural Computation(Springer, 2019) pp. 232–256

2019
[15]

E. M. H. E. B. Ekanayake and N. Shukla, Different paths, same destination: Designing physics-inspired dynamical systems with engineered stability to minimize the Ising hamiltonian, Physical Review Applied24, 024008 (2025)

2025
[16]

J. G. Restrepo, E. Ott, and B. R. Hunt, Onset of synchro- nization in large networks of coupled oscillators, Physical Review E71, 036151 (2005)

2005
[17]

E. C. Stoner and E. P. Wohlfarth, A mechanism of mag- netic hysteresis in heterogeneous alloys, Philosophical Transactions of the Royal Society of London. Series A, Mathematical and Physical Sciences240, 599 (1948)

1948
[18]

Tannous and J

C. Tannous and J. Gieraltowski, The Stoner–Wohlfarth model of ferromagnetism, European Journal of Physics 29, 475 (2008)

2008
[19]

Braun, Exact activation energy of magnetic single do- main particles, Journal of Magnetism and Magnetic Ma- terials283, 1 (2004)

D. Braun, Exact activation energy of magnetic single do- main particles, Journal of Magnetism and Magnetic Ma- terials283, 1 (2004)

2004

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.

[1] [1]

Mohseni, P

N. Mohseni, P. L. McMahon, and T. Byrnes, Ising ma- chines as hardware solvers of combinatorial optimization problems, Nature Reviews Physics4, 363 (2022)

2022

[2] [2]

Zhang, Q

T. Zhang, Q. Tao, B. Liu, A. Grimaldi, E. Raimondo, M. Jim´ enez, M. J. Avedillo, J. Nunez, B. Linares- Barranco, T. Serrano-Gotarredona, G. Finocchio, and J. Han, A review of Ising machines implemented in con- ventional and emerging technologies, IEEE Transactions on Nanotechnology23, 704 (2024)

2024

[3] [3]

Lucas, Ising formulations of many NP problems, Fron- tiers in Physics2, 74887 (2014)

A. Lucas, Ising formulations of many NP problems, Fron- tiers in Physics2, 74887 (2014)

2014

[4] [4]

Chang, C.-F

Y.-J. Chang, C.-F. Nien, K.-P. Huang, Y.-T. Zhang, C.-H. Cho, and C.-R. Chang, Quantum computing for optimization with Ising machine, IEEE Nanotechnology Magazine18, 15 (2024)

2024

[5] [5]

P. L. McMahon, A. Marandi, Y. Haribara, R. Hamerly, C. Langrock, S. Tamate, T. Inagaki, H. Takesue, S. Ut- sunomiya, K. Aihara,et al., A fully programmable 100- spin coherent Ising machine with all-to-all connections, Science354, 614 (2016)

2016

[6] [6]

Ezawa, E

M. Ezawa, E. Lebrasseur, and Y. Mita, Ising machine based on bistable microelectromechanical systems, Jour- nal of the Physical Society of Japan91, 114601 (2022)

2022

[7] [7]

Grimaldi, L

A. Grimaldi, L. Mazza, E. Raimondo, P. Tullo, D. Ro- drigues, K. Y. Camsari, V. Crupi, M. Carpentieri, V. Puliafito, and G. Finocchio, Evaluating spintronics- compatible implementations of Ising machines, Physical Review Applied20, 024005 (2023)

2023

[8] [8]

J. Chou, S. Bramhavar, S. Ghosh, and W. Herzog, Analog coupled oscillator based weighted Ising machine, Scien- tific Reports9, 14786 (2019)

2019