SK Reader

Full support and positive density in the zero-temperature SK model

Will Cook · An account of the quantitative proof of OpenAI (2026a) · 3 November 2026

AI assistance. Will Cook directed the work. AI agents based on large language models produced and revised this manuscript.

Abstract. For the pure, zero-field Sherrington–Kirkpatrick model at zero temperature, the Stieltjes measure of every integrable minimiser of the Parisi functional has full support on [0,1)[0,1) and a smooth density that is strictly positive on (0,1)(0,1). We reconstruct OpenAI’s quantitative proof, which uses one spatial-moment inequality twice. On a positive constant interval, the inequality makes the derivative of the gradient second moment strictly convex, contradicting the endpoint conditions imposed by minimality, and together with the bounded gradient it also excludes terminal gaps. Once the support is full, moment identities give smoothness, and a local constant replacement transfers the inequality to the minimiser and forces its density to be positive. H.-B. Chen previously proved full support and smooth nonnegative density. We isolate the fixed-time spatial moment and the positive margin that survives the local replacement, and we import its shape comparisons and full polynomial identity. Along the way we derive the scalar value identity and explain its classical optimisation consequence. For the quantum approximate optimization algorithm (QAOA), we identify the rooted energy of a finite Gaussian observable and the separate circuit theorem that approximates it, and a further second-moment theorem then gives a measured near-optimal configuration on the same random instance, with the circuit word fixed before the number of spins grows. The construction gives no bound on the depth needed for a given accuracy.

1. The model and the result

For independent standard Gaussian couplings JijJ_{ij}, the pure, zero-field Sherrington–Kirkpatrick (SK) Hamiltonian is Hn(σ)=n−1/2∑i<jJijσiσj,σ∈{−1,1}n.H_n(\sigma)=n^{-1/2}\sum_{i<j}J_{ij}\sigma_i\sigma_j, \qquad \sigma\in\{-1,1\}^n. With the maximisation convention, its limiting ground-state energy per spin is P∗:=lim⁡n→∞n−1Emax⁡σHn(σ).P_*:=\lim_{n\to\infty}n^{-1}\mathbb E\max_\sigma H_n(\sigma). The Parisi formula expresses P∗P_* as a minimum over nondecreasing functions. We study the minimiser itself. Its increments charge every interval of overlap scales below one, and they have a strictly positive density at each interior scale. One quantitative bound gives both facts: it first excludes constant intervals and then, after an approximation, gives pointwise positivity.

The overlap of two configurations is R(σ,τ)=n−1∑iσiτiR(\sigma,\tau)=n^{-1}\sum_i\sigma_i\tau_i, and it determines the energy covariance, E[Hn(σ)Hn(τ)]=nR2/2−1/2\mathbb E[H_n(\sigma)H_n(\tau)]=nR^2/2-1/2. Hence the Gaussian law of the energies is encoded by the overlaps. We use the covariance convention ξ(t)=t2/2\xi(t)=t^2/2, so ξ′′=1\xi''=1. Adding one independent centred Gaussian of variance 1/21/2 to every configuration’s energy gives covariance nξ(R)n\xi(R) and leaves the expected maximum unchanged, which is the normalisation of the ground-state Parisi formula of Auffinger and Chen (2017), Theorem 1. The model has no higher-order interactions and no external field.

Let U\mathcal U consist of the nonnegative, nondecreasing, right-continuous functions γ:[0,1)→[0,∞)\gamma:[0,1)\to[0,\infty) satisfying ∫01γ(t) dt<∞\int_0^1\gamma(t)\,dt<\infty, and define the locally finite Stieltjes measure by μγ([0,t])=γ(t),0≤t<1.\mu_\gamma([0,t])=\gamma(t),\qquad 0\leq t<1. Thus μγ((s,t])=γ(t)−γ(s)\mu_\gamma((s,t])=\gamma(t)-\gamma(s), so the measure charges each interval with the increment of γ\gamma across it. It may have a point mass (an atom) of size γ(0)\gamma(0) at zero, and its total mass need not be finite or equal to one. Support statements refer to the relative topology of [0,1)[0,1). A support gap is a relatively open interval of positive length and zero mass, and the cumulative coefficient is constant on it.

The Parisi functional P(γ)\mathcal P(\gamma) is defined through an auxiliary value function Φγ(t,x)\Phi_\gamma(t,x), in which x∈Rx\in\mathbb R is a spatial variable and the overlap parameter tt plays the role of time. Write Φ=Φγ\Phi=\Phi_\gamma and use subscripts for partial derivatives. The backward equation, terminal condition and functional are Φt=−12(Φxx+γ(t)Φx2),Φ(1,x)=∣x∣,P(γ)=Φγ(0,0)−12∫01tγ(t) dt.\Phi_t=-\tfrac12\bigl(\Phi_{xx}+\gamma(t)\Phi_x^2\bigr), \qquad \Phi(1,x)=|x|, \qquad \mathcal P(\gamma)=\Phi_\gamma(0,0)-\tfrac12\int_0^1t\gamma(t)\,dt. The terminal value ∣x∣|x| is the zero-temperature value of one spin in a field xx. For an integrable coefficient, the solution of the PDE is defined by approximation from bounded monotone coefficients. OpenAI (2026a), §1 and Lemmas 2.1–2.2 give the exact control definition and the stability theorem.

To turn minimality of P\mathcal P into a condition on γ\gamma, we use a random process whose gradient moment measures the first-order cost of changing γ\gamma. Write u=Φxu=\Phi_x for the spatial gradient and a=Φxxa=\Phi_{xx} for its spatial derivative. Let BB be standard Brownian motion, and define the diffusion XX and its self-consistency function qq by dXt=γ(t)u(t,Xt) dt+dBt,X0=0,q(t)=E[u(t,Xt)2].dX_t=\gamma(t)u(t,X_t)\,dt+dB_t,\qquad X_0=0, \qquad q(t)=\mathbb E[u(t,X_t)^2]. The discrepancy q(t)−tq(t)-t enters the first variation of P\mathcal P, and minimality will force q(t)=tq(t)=t on the support of μγ\mu_\gamma. The SDE has a unique strong solution. Unless an endpoint limit is justified explicitly, all stochastic calculations below take place on compact time intervals ending before one.