| Gottesman–Kitaev–Preskill (GKP) code | |
|---|---|
| Name | Gottesman–Kitaev–Preskill (GKP) code |
| Developer | Daniel Gottesman, Alexei Kitaev, John Preskill |
| Introduced | 2001 |
| Field | Quantum error correction |
| Related | Continuous-variable quantum information, Quantum computing, Bosonic code |
Gottesman–Kitaev–Preskill (GKP) code
The Gottesman–Kitaev–Preskill (GKP) code is a bosonic quantum error correction code that encodes a logical qubit into the continuous-variable degrees of freedom of an harmonic oscillator mode. It is important because it offers a path to protect quantum information against small displacement errors in phase space, enabling fault-tolerant operations in architectures that use photonic quantum computing or superconducting microwave resonators.
The GKP code was proposed in 2001 by Daniel Gottesman, Alexei Kitaev, and John Preskill to leverage continuous-variable systems for digital quantum information processing. It built on earlier work in quantum error correction and stabilizer code theory, and connected discrete qubit encodings to continuous-variable techniques used in quantum optics and optomechanics. The proposal influenced subsequent bosonic-code research such as the cat code and binomial code, and prompted experimental efforts in platforms including trapped ion motion modes, optical modes, and circuit quantum electrodynamics (cQED).
The GKP logical states are idealized comb-like superpositions of position eigenstates (|q⟩) with periodic spacing in phase space, forming eigenstates of two commuting displacement operators (stabilizers). Logical Pauli operators correspond to half-period displacements. In the ideal limit, the logical |0_L⟩ and |1_L⟩ are infinite-energy Dirac combs; physical realizations use finitely squeezed approximations described by Gaussian envelopes. The code is naturally expressed using the Wigner quasiprobability distribution and phase-space operators: position q and momentum p quadratures, Weyl displacement operators, and the Heisenberg–Weyl group. The GKP stabilizer group is an abelian subgroup of the oscillator's displacement group, making the code a continuous-variable counterpart of stabilizer formalism used for Calderbank–Shor–Steane (CSS) codes.
The dominant errors addressed by the GKP code are small random shifts (displacements) in q and p due to noise sources like photon loss, thermal excitation, or classical control imprecision. Error correction proceeds by measuring modular quadratures (q mod √π, p mod √π) using ancilla-assisted homodyne-like protocols to extract syndrome information without collapsing the logical state. Recovery applies corrective displacements conditioned on syndromes; for finite-energy GKP states this is approximate and must trade off between syndrome precision and added noise. Analysis of logical error rates uses models such as the Gaussian displacement channel and photon-loss channel, often leveraging quantum capacity and fidelity metrics.
Experimental approaches aim to generate approximate GKP states and perform syndrome extraction. Notable platforms include optical implementations using squeezed states and linear optics with ancillary non-Gaussian resources, microwave cavity modes coupled to superconducting qubits in cQED experiments, and motional modes of trapped ions. Groups at institutions like Yale University, Caltech, ETH Zurich, and companies such as Google (company) and IBM have reported progress in preparing grid states, performing logical operations, and demonstrating basic error correction primitives. Key experimental techniques include reservoir engineering, gate-based synthesis with non-Gaussian ancillae, and teleportation-based state preparation. Demonstrations typically report finite squeezing (measured in dB) and characterize state quality via Wigner tomography and logical error benchmarking.
The GKP code can serve as a hardware-efficient substrate for fault-tolerant quantum computing by concatenation with discrete qubit codes such as the surface code or Steane code. Encoding physical oscillators as GKP qubits allows continuous-variable noise to be converted into discrete Pauli error channels that higher-level topological quantum error correction can handle. Threshold analyses combine the finite-squeezing noise model with thresholds of concatenated codes; proposals show that realistic levels of squeezing and cavity coherence could enable surface-code thresholds with reduced overhead compared to pure qubit hardware. Fault-tolerant GKP operations exploit Gaussian gates, Clifford gates, and magic-state injection for universal computation.
GKP codes are proposed for fault-tolerant quantum processors, quantum memories in hybrid architectures, and robust quantum communication over bosonic channels such as free-space or fiber links. Their capacity to correct small shifts makes them suitable for continuous-variable quantum key distribution variants and for improving transduction between microwave and optical domains. Additionally, GKP encodings facilitate efficient implementation of logical gates via Gaussian operations and feedforward, and they play a role in proposals for scalable modular quantum networks and error-protected quantum repeaters.
Practical deployment of GKP codes faces challenges: creating high-quality, finite-energy GKP states with sufficient squeezing and fidelity; implementing nondestructive, low-noise modular quadrature measurements; mitigating photon loss and non-Gaussian noise; and integrating GKP elements into scalable hardware with error-corrected logical gates. Open theoretical questions include rigorous fault-tolerance thresholds in realistic noise models, optimal decoding algorithms for concatenated architectures, resource-efficient state preparation, and new hybrid codes combining GKP with bosonic and topological schemes. Continued experimental progress and cross-disciplinary collaboration among quantum optics, condensed matter physics, and quantum information science communities are central to addressing these gaps.
Category:Quantum error correction Category:Quantum information theory Category:Continuous-variable quantum information