⚛️ quantum physics

Parisi Formula for the ground state energy of quantum p-Spin Hamiltonians

This paper establishes that the ground state energy of quantum $p$ -local spin glass Hamiltonians converges to a Parisi-type variational formula for all $p \ge 2$ , thereby resolving a question left open by Anschuetz et al. (2025) and demonstrating the universality of this limit across non-Gaussian interactions.

Original authors: Sohom Bhattacharya

Published 2026-09-18

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Sohom Bhattacharya

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

In the vast landscape of quantum physics, researchers often study systems where countless tiny particles interact in ways that seem chaotic and unpredictable. These are known as spin glass models, a concept borrowed from the study of magnetic materials where the internal forces between atoms are so tangled that the material gets stuck in a state of confusion, unable to settle into a simple, orderly pattern. When scientists add the rules of quantum mechanics to this mix, the problem becomes even more complex. Instead of just being magnetic or not, these quantum particles can exist in a superposition of states, and their interactions create a landscape of possible energies that is incredibly rugged. The central question for physicists in this specific context is to find the "maximal energy" achievable by a specific type of quantum state. Knowing this value is crucial because it tells us the fundamental limits of what these systems can achieve under certain constraints. However, calculating this highest energy for a quantum system with many interacting parts is notoriously difficult, often requiring approximations that might miss the true answer.

A researcher has now provided a precise mathematical description of the maximum energy achievable by a specific, simplified type of quantum state in these complex systems. They focused on a class of quantum models called p-spin Hamiltonians, which describe how groups of quantum bits, or qubits, interact with one another. In these models, the interactions are random, much like a system where every particle is connected to every other particle by a force that is determined by a roll of the dice. The researcher was interested in finding the highest energy that can be reached if the system is restricted to "product states." In plain terms, a product state is a configuration where every single qubit acts independently, without being entangled with the others. While entangled states are the most powerful resource in quantum computing, product states are much easier to work with and understand. The researcher proved that as the number of qubits grows very large, the maximum energy achievable by these independent states settles down to a specific, predictable value.

The author demonstrated that this limiting value is not a random number but is determined by a sophisticated mathematical recipe known as a Parisi-type variational formula. This formula acts like a map, guiding researchers to the exact peak energy by optimizing a specific function over a range of possibilities. Before this work, it was known that such a limit existed for certain types of these systems, but no one had been able to write down the exact formula for every case. The researcher filled this gap, showing that for any number of interacting particles greater than or equal to two, this limit exists and can be calculated. They also showed that this result is robust; it does not depend on the specific type of randomness used to generate the interactions. Whether the random forces follow a standard bell-curve distribution or a different, more unusual pattern, the final energy limit remains the same. This finding confirms that the behavior of these systems is universal, governed by deep structural principles rather than the specific details of the noise affecting them.

To reach this conclusion, the researcher translated the difficult quantum problem into a more manageable form. They used a geometric representation where each qubit is described by a point on a sphere, turning the quantum search for the best state into a problem of finding the highest point on a vast, multi-dimensional landscape. By comparing this landscape to a known mathematical model of a glassy system, they were able to apply powerful tools from statistical physics. They proved that the complex quantum optimization problem could be reduced to a simpler, scalar problem that could be solved using the Parisi formula. This approach allowed them to bypass the need for simulating the entire quantum system, which would be impossible for large numbers of particles, and instead focus on the underlying mathematical structure that dictates the outcome.

The paper also addressed how this maximum energy behaves when the number of interacting particles becomes extremely large. The researcher showed that their new formula correctly predicts the energy growth rate in this extreme limit, matching results that were previously known only for very large systems. This serves as a consistency check, confirming that their general formula works not just for small systems but also scales correctly to the infinite limit. Furthermore, they proved that the result holds true even if the random interactions are not perfectly Gaussian, a common assumption in physics. By showing that the energy limit stays the same for a broad class of non-Gaussian interactions, they established that the phenomenon is a fundamental property of the system's structure, not an artifact of a specific mathematical choice.

This work resolves a question that had remained open in the field, providing a complete and explicit characterization of the maximum energy for product states in these quantum models. It moves the field from a state of partial understanding, where limits were known to exist but not defined, to a state of precise calculation. The findings suggest that even in the chaotic world of quantum spin glasses, there are rigid mathematical laws that determine the boundaries of what is possible. By identifying these laws, the researcher has provided a new tool for understanding the limits of quantum systems, offering a clear path for future investigations into how quantum matter organizes itself under complex, random forces. The result is a definitive answer to a long-standing puzzle, grounded in rigorous proof and applicable across a wide range of physical scenarios.

Technical Summary: Parisi Formula for the Ground State Energy of Quantum p-Spin Hamiltonians

Problem Statement
The paper addresses the problem of estimating the ground state energy of quantum $p$ -local spin glass Hamiltonians, which serve as quantum analogues to classical spin glass models. Specifically, the author focuses on the maximum energy achievable by product states (unentangled states) in a system of $n$ qubits. The Hamiltonian $H_{n,p}$ is defined as a sum of $p$ -local Pauli string interactions with independent standard Gaussian coefficients. While the limit of the maximal energy for classical spin glasses is well-characterized by Parisi-type variational formulas, the quantum analogue for fixed $p$ remained an open question. Previous work by [AGK25] established the existence of the limit for even $p$ and derived large- $p$ asymptotics, but did not provide an explicit characterization for fixed $p \ge 2$ . This paper aims to fill that gap by deriving a Parisi-type variational formula for the limiting maximal product-state energy for any integer $p \ge 2$ .

Methodology
The author employs a multi-step reduction and comparison strategy to derive the limiting energy:

Bloch-Sphere Reduction: The optimization over the set of product states $S_n^{\text{prod}}$ is reduced to a classical optimization problem on the product of Bloch spheres $(S^2)^n$ . Using the Bloch vector representation of single-qubit states, the quantum expectation value $\langle \phi | H_{n,p} | \phi \rangle$ is mapped to a classical random process $X_{n,p}(s)$ defined on vectors $s \in (S^2)^n$ .
Gaussian Process Approximation: The author constructs a centered Gaussian process $Y_{n,p}$ on $(S^2)^n$ with a specific covariance structure $n(q(s,t))^p$ , where $q(s,t)$ is the overlap between configurations. They prove that the free energy (and consequently the ground state energy) of the original process $X_{n,p}$ converges to that of $Y_{n,p}$ as $n \to \infty$ , utilizing Gaussian concentration and interpolation arguments.
Finite-Temperature Parisi Formula: The problem is then linked to the free energy of a vector spin-glass model. By invoking recent results on vector spin glasses [CM25], the author establishes a finite-temperature Parisi formula involving a path of $3 \times 3$ positive semidefinite matrices.
Symmetry Reduction: Leveraging the rotational symmetry of the model (invariance under orthogonal transformations $O(3)$ ), the matrix-valued variational problem is shown to reduce to a scalar problem. The optimal path is identified as a scalar multiple of the identity matrix, simplifying the functional to depend on a single non-decreasing function $\alpha$ .
Zero-Temperature Limit: To obtain the ground state energy, the author takes the zero-temperature limit ( $\beta \to \infty$ ) of the finite-temperature formula. This involves adapting arguments from [AC17] to handle the stochastic control representation of the Parisi partial differential equation (PDE) in the limit, ultimately yielding a variational formula over non-decreasing functions $\gamma$ .
Universality: A separate proof using the Lindeberg comparison principle and Efron-Stein inequality is provided to show that the limiting energy is universal, holding for a broad class of non-Gaussian interaction coefficients satisfying specific moment conditions.

Key Contributions and Results

Existence and Variational Formula (Theorem 1): The paper proves that for every integer $p \ge 2$ , the limit of the expected maximal product-state energy $E_{\text{prod}}^*(p) = \lim_{n \to \infty} \mathbb{E}[\max_{|\phi\rangle \in S_n^{\text{prod}}} \langle \phi | H_{n,p} | \phi \rangle / \sqrt{n}]$ exists. This limit is given explicitly by a Parisi-type variational formula:
$E_{\text{prod}}^*(p) = \inf_{\gamma \in \mathcal{U}} P_p(\gamma)$
where $\mathcal{U}$ is the space of non-negative, non-decreasing, right-continuous functions, and $P_p(\gamma)$ is defined via a stochastic control problem involving a Brownian motion in $\mathbb{R}^3$ . The infimum is attained.
Large- $p$ Asymptotics (Proposition 1): As a consequence of the variational formula, the author recovers the known large- $p$ asymptotics, proving that $\lim_{p \to \infty} E_{\text{prod}}^*(p) / \sqrt{2 \log p} = 1$ . This provides an alternative proof to the result in [AGK25].
Universality (Proposition 2): The paper establishes that the limiting maximal energy is universal. It remains unchanged if the Gaussian coefficients are replaced by independent coefficients with zero mean, unit variance, and satisfying a Lindeberg-type condition (specifically, the tail probability of large coefficients decays sufficiently fast). This extends universality results previously known for the full spectrum of pure states to the specific class of product states.

Significance and Claims
The paper claims to settle a specific open question left by [AGK25] regarding the explicit characterization of the fixed- $p$ limit for quantum $p$ -spin Hamiltonians restricted to product states. By providing a Parisi-type variational formula, the work offers a tractable method to compute the ground state energy for any $p \ge 2$ , bridging the gap between the classical spin glass theory and its quantum product-state analogue.

The author notes that while the formula characterizes the energy, they do not address the computational complexity of finding a product state that achieves this energy in polynomial time, referencing [AGG26] for lower bounds on such circuit complexity. Furthermore, the paper distinguishes its focus on product states from other random quantum Hamiltonian models (like the SYK model or quantum Sherrington-Kirkpatrick model) where entangled states play a central role, noting that the techniques used for general pure states (e.g., approximating the maximum via the log-partition function) do not directly apply to the product-state constraint.

The results are presented as rigorous mathematical derivations within the fields of probability theory and statistical mechanics, relying on established tools such as the Ruelle probability cascade, Gaussian concentration, and stochastic control representations of PDEs.

More like this