Speaker
Description
The Gottesman-Kitaev-Preskill (GKP) code is a bosonic encoding for fault-tolerant quantum computing that protects a qubit using the phase space of a harmonic oscillator. A key feature of the code is that small continuous displacement errors can be converted into effective qubit-level Pauli errors. In ideal GKP error correction, the measured syndrome is usually decoded by standard binning: the syndrome is assigned to the nearest GKP lattice point, and the parity of that point determines the Pauli correction. This rule is natural in ideal symmetric settings, but realistic implementations use finite-energy GKP states and are affected by Gaussian operations and physical noise. These effects can change the syndrome distribution, so the nearest lattice point may not give the correction that best preserves the logical qubit.
This work investigates how the GKP decoder can be tailored to the physical circuit. Instead of using only nearest-lattice-cell geometry, I compare candidate Pauli corrections using syndrome-dependent scores derived from the noisy error-correction process. These scores are used to construct alternative decoding rules and to study when they agree with, or deviate from, standard binning.
The performance of a decoder is assessed through the effective logical channel induced by the GKP error-correction step. In particular, I consider how different correction rules change the logical Pauli error probabilities and the channel fidelity relative to the ideal logical identity operation. This provides a practical way to ask whether a decoder is not only locally plausible from the measured syndrome, but also optimal for preserving encoded information. Understanding how finite-energy effects and Gaussian noise alter the best correction rule is a step toward designing inner GKP decoders for larger concatenated GKP-qubit architectures.
| I am the presenting author | Yes |
|---|