⚛️ quantum physics

Continuous-variable designs and design-based shadow tomography from random lattices

This paper demonstrates that Gottesman-Kitaev-Preskill (GKP) code states form a rigged continuous-variable 2-design, which is leveraged to construct a practical shadow tomography protocol with derived sample complexity bounds and physical implementation schemes for continuous-variable quantum systems.

Original authors: Jonathan Conrad, Joseph T. Iosue, Ansgar G. Burchards, Victor V. Albert

Published 2026-10-09

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Jonathan Conrad, Joseph T. Iosue, Ansgar G. Burchards, Victor V. Albert

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

Quantum mechanics describes a world that is continuous, like the smooth flow of a river, and this continuity is fundamental to nature. Many natural systems, such as light waves or the vibrations of atoms, are inherently continuous. To harness these systems for powerful computing, scientists need a way to measure and understand them without getting lost in an ocean of infinite possibilities. This is where the concept of a "design" becomes essential. In mathematics and science, a design is a carefully selected set of samples that acts as a skeleton for a much larger space. Instead of measuring every single point in a vast landscape, a good design allows you to take a few strategic measurements that accurately represent the whole. For decades, scientists have used these designs to study discrete quantum systems, but extending this powerful tool to continuous systems has remained a difficult, unsolved puzzle.

A team of researchers has now solved this puzzle by turning to a specific type of quantum state known as a lattice state. These are not ordinary waves but highly structured patterns of light or vibration that repeat in a grid-like fashion, much like the points on a graph paper. These states are the building blocks of a famous error-correction method called Gottesman-Kitaev-Preskill, or GKP, which helps protect fragile quantum information from noise. The researchers discovered that if you take all possible versions of these lattice states and shift them around in every possible way, they form a rigorous mathematical design for continuous systems. This means that by randomly selecting and measuring these specific states, scientists can reconstruct the full picture of a complex quantum system with high efficiency. It is a breakthrough that bridges the gap between the theoretical tools used for discrete computers and the practical needs of continuous ones.

The core of this discovery lies in proving that these lattice states form what is called a "two-design." In simpler terms, this means that if you average the results of measurements taken from this specific collection of states, the result is mathematically identical to what you would get if you could somehow measure every single possible state in the universe. This property is crucial for a technique called shadow tomography, which is a method for learning about a quantum system by taking many small, random snapshots. Previously, shadow tomography was a powerful tool for discrete qubit systems, but it lacked a reliable version for continuous systems. By proving that GKP lattice states form this design, the authors have provided the missing key. They showed that these states can be used to build a protocol that extracts information about a continuous system just as effectively as the best protocols do for discrete ones.

The researchers did not stop at the theoretical proof; they also worked out how to actually perform these measurements in a real laboratory. They described specific physical setups, or "gadgets," that use standard quantum equipment to interact with these lattice states. One method involves using a standard qubit to help measure the continuous system, while another uses an auxiliary lattice state to perform the measurement directly. These are not just abstract ideas; the paper details the circuits and steps required to implement them, showing that the necessary technology already exists or is within reach. This turns a mathematical curiosity into a practical recipe for experimentalists who want to characterize complex quantum states in optical or mechanical systems.

A significant part of the work involved calculating how many measurements are needed to get a good answer, a concept known as sample complexity. The team derived strict mathematical bounds showing that the number of samples required grows in a manageable way as the system gets larger. They found that for global measurements, where the whole system is treated as one unit, the efficiency depends on how quickly the system's properties fade out in phase space. For local measurements, where parts of the system are measured independently, the efficiency depends on how entangled or connected the different parts are. Interestingly, their results suggest that the difficulty of measuring these systems is not determined by how "local" or "global" the observable is, but rather by how much the system's information is spread out across different parts. This is a distinct difference from how discrete systems behave, offering new insights into the nature of continuous quantum information.

The study also addressed a major hurdle: the fact that the ideal lattice states described in the math are infinite in energy and cannot exist in reality. The researchers showed that by using slightly "squeezed" or regularized versions of these states—states that are physically realizable and have finite energy—the design property still holds approximately. This means the protocol is robust enough to work with the imperfect, real-world states that experimentalists can actually create. They also explored a conjecture regarding the energy limits of these states, suggesting that if the system's energy is bounded, the efficiency of the measurement protocol remains predictable and efficient. While they could not prove whether these states form an even stronger type of design (a 3-design), the evidence they provided for the two-design property is solid and sufficient for the proposed tomography protocol.

Ultimately, this work elevates the status of GKP lattice states from a tool for error correction to a fundamental resource for quantum measurement. It demonstrates that the same states used to fix errors in quantum computers can also be used to efficiently map out the state of those computers. By establishing a rigorous link between lattice theory and quantum information, the authors have opened a new path for characterizing continuous-variable systems. This paves the way for more accurate simulations of molecular vibrations, better sensors for detecting gravitational waves, and more efficient quantum computers that operate natively in the continuous domain. The paper provides both the theoretical foundation and the practical instructions needed to bring these advanced measurement techniques into the lab, marking a significant step forward in our ability to understand and control the continuous quantum world.

Technical Summary: Continuous-variable designs and design-based shadow tomography from random lattices

Problem Statement
Quantum information processing is increasingly recognizing continuous-variable (CV) systems (e.g., photonic or phononic modes) as a promising strategy for scalable quantum computing. However, extracting information from these infinite-dimensional systems requires robust protocols analogous to those available for qubits. A critical tool in qubit-based quantum information is the concept of designs—specifically $t$ -designs—which provide a finite or structured subset of states that mimic the statistical properties of the full Haar-random ensemble up to the $t$ -th moment. While finite-dimensional designs are well-established and underpin protocols like shadow tomography, their extension to infinite-dimensional CV spaces has been limited. Previous work established CV 2-designs for functions on the real line ( $d=1$ ), but the existence of natural designs for higher-dimensional CV systems ( $d > 1$ ) remained an open question. Furthermore, there was a need for practical shadow tomography protocols for CV systems that could leverage existing quantum error correction resources.

Methodology
The authors investigate the properties of Gottesman-Kitaev-Preskill (GKP) states, which are quantum lattice states defined as infinite superpositions of coherent states at points within a symplectic lattice in phase space. The core methodology involves:

Mathematical Construction: The authors define an ensemble $\mathcal{X}_n$ consisting of all displaced GKP states for all possible $2n$ -dimensional symplectic lattices. This ensemble is parameterized by the space of symplectic lattices $Y_n = \text{Sp}_{2n}(\mathbb{Z}) \setminus \text{Sp}_{2n}(\mathbb{R})$ and the fundamental domain of each lattice.
Lattice Theory Tools: The proof relies on two key tools from lattice theory:
- The Symplectic Mean Value Formula: A theorem allowing the conversion of integrals over the space of symplectic lattices into integrals over the entire phase space.
- Second-Moment Bounds: Recent bounds on functions over the space of symplectic lattices used to analyze variance.
Design Verification: The authors verify that the ensemble $\mathcal{X}_n$ satisfies the definition of a rigged 2-design. This involves showing that the ensemble resolves permutation-symmetric subspaces, satisfying the condition $\int d\mu(X) |X\rangle\langle X|^{\otimes t'} = a_{t'} \Pi_{t'}$ for $t' \leq 2$ , where $\Pi_{t'}$ is the projector onto the symmetric subspace.
Shadow Tomography Protocols: Using the 2-design property, the authors construct two shadow tomography protocols:
- Global GKP Shadows: Analogous to global Clifford shadows for qubits, utilizing the full ensemble of $n$ -mode GKP states.
- Local GKP Shadows: Analogous to local Pauli shadows, utilizing tensor products of single-mode GKP states.
Physical Implementation: The paper details physical circuits to implement these protocols, including strategies using qubit-assisted phase estimation and auxiliary GKP modes for continuous-valued phase estimation.

Key Contributions and Results

Existence of CV 2-Designs: The primary theoretical result is Theorem 1, which proves that the ensemble of all displaced GKP states for all symplectic lattices forms a rigged continuous-variable 2-design for any $n$ -mode system. This answers the open question regarding the existence of natural designs for higher-dimensional CV spaces.
Sample Complexity Bounds:
- Global Shadows: The authors derive sample complexity bounds for global GKP shadows (Theorem 2). The variance of the estimator depends on the trace of the squared observable ( $\text{Tr}[O^2]$ ) and the $L_1$ norm of the observable's characteristic function. The $L_1$ norm serves as a measure of non-Gaussianity, a feature absent in qubit bounds.
- Local Shadows: For local GKP shadows, the authors derive bounds (Theorem 3) under "physicality" assumptions: the input state has a bounded average photon number and an exponentially decaying characteristic function. Crucially, they propose a conjecture (Eq. 14) regarding a lower bound on an effective squeezing parameter scaling with the inverse average photon number. Under this conjecture, the sample complexity scales similarly to qubit local shadows but depends on the "inseparability" (total purity across bipartitions) of the observable rather than its strict tensor-product locality.
Implementation Recipes: The paper provides concrete circuit diagrams and strategies for implementing the shadow tomography protocol. This includes:
- Qubit-assisted methods: Using dominant-bit phase estimation to extract GKP syndrome information.
- GKP-assisted methods: Using a single auxiliary GKP mode to measure multimode stabilizers via continuous-valued phase estimation, a novel strategy introduced in this work.
- SWAP Tests: Noting that direct estimation of probabilities $\langle X|\rho|X\rangle$ is possible via SWAP tests between resource states and the input.

Significance and Claims
The paper positions GKP states not merely as error-correcting codes but as a powerful tomographic resource for continuous-variable systems, placing them on par with Clifford circuits and Pauli measurements in qubit systems.

Bridging Theory and Experiment: The work connects abstract lattice theory with practical quantum information tasks. It demonstrates that GKP codes, which have already been realized in several experiments (e.g., trapped ions, superconducting cavities), can be directly utilized for efficient state characterization.
Non-Gaussianity as a Resource: The derived bounds highlight the role of the $L_1$ norm of the characteristic function, linking the efficiency of the protocol to the non-Gaussian nature of the observables and states. This offers a new perspective on the "cost" of tomography in CV systems.
Limitations and Modesty: The authors are explicit about the limitations of their results. They do not prove that GKP states form a 3-design (which would yield tighter bounds independent of characteristic functions), though they note that their ensemble possesses "just enough structure" to yield useful 2-design bounds. The local shadow bounds rely on a conjecture regarding the squeezing parameter of physical states. The paper frames these results as a foundational step, suggesting that the tools developed (mean value formulas, second-moment bounds) may have broader applications in quantum information and cryptography (e.g., learning with errors problems).

In summary, the paper establishes a rigorous mathematical framework for using random GKP states as a 2-design, enabling the construction of efficient shadow tomography protocols for continuous-variable quantum systems and providing the physical blueprints for their experimental realization.

Technical Summary: Continuous-variable designs and design-based shadow tomography from random lattices

More like this