arXiv:2609.39758
Theory of spacetime quantum fault tolerance
Yijia Xu, Yixu Wang, Zi-Wen Liu
Abstract
We present a unified algebraic theory of Clifford spacetime quantum fault tolerance, which underpins universal fault-tolerant quantum computation. We construct a spacetime chain complex from a tensor network representation of a circuit that encodes fault propagation, detector syndromes, and logical actions through the associated gauge and stabilizer structure. We show that this theory (1) unifies various formulations of quantum fault tolerance including spacetime codes, Gottesman's gadget framework, fault complexes, and ZX calculus; (2) connects spacetime distances to the gadget-level correctness properties in Gottesman's formalism; (3) naturally accommodates a range of fault models, in particular phenomenological and circuit-level models, bringing these often separately formulated descriptions into a common algebraic framework through different tensor decompositions and fault assignments; and (4) recovers many known fault tolerance constructions and results as special cases, including repeated syndrome extraction, correlated detectors for transversal and fold-transversal gates, Steane- and Knill-type syndrome extraction gadgets, and single-shot error correction.
AI Summary
The paper asks how algebraic descriptions of noisy Clifford circuits relate to operational gadget-correctness conditions, and how spacetime codes, circuit detectors, fault complexes, and ZX descriptions can be translated into one another. It represents circuit components by stabilizer tensors and retains internal tensor legs as spacetime fault locations. Tensor relations, wire contractions, input stabilizers, and measurements generate a subsystem-code gauge group $\mathcal G$; its center supplies spacetime stabilizers. These define a chain complex with maps $G^{\mathsf T}$ and $S\Omega$, encoding fault propagation as gauge equivalence and logical faults as homology classes. A detector-restricted bulk complex distinguishes measured constraints from boundary stabilizer information. For an input code of distance $d=2t+1$, the paper establishes that correctness against $t$ faults implies spacetime distance $d_{\mathrm{st}}=d$, and claims equivalence for circuits without Pauli feedback. For syndrome extraction, it gives sufficient correctness conditions involving $d_{\mathrm{st}}\geq2t+1$, a decoder-dependent bulk effective distance, and bounded fault amplification. Examples recover repeated extraction, correlated detectors across transversal and fold-transversal gates, and single-shot bounds from soundness. Explicit calculations give spacetime subsystem-code parameters $[\![11n,k,9n+k,d]\!]$ for Steane extraction and $[\![9n,k,7n+k,d]\!]$ for Knill extraction. CSS tensor networks are related to fault complexes through chain maps. The framework concerns Clifford protocols and specified Pauli-fault assignments; several examples use phenomenological noise and minimum-weight recovery. It does not establish positive noise thresholds: probabilistic noise bounds, efficient decoding, and robustness under repeated gadget composition remain future directions.
Central claim may not hold: The defined bulk effective distance is a minimum failing weight, but subsequent results use it as a guaranteed correction radius, and the claimed inequality contradicts the minimum-weight-recovery bound d_st <= 2d_bulk. The single-shot section claims an exact bulk effective distance from soundness estimates that establish only a sufficient correctability range. The converse of the no-feedback correctness-equivalence theorem does not justify replacing a suitable spacetime recovery with the perfect output-code recovery required by the correctness property. Lemma outputequiv incorrectly excludes propagated input stabilizers from the difference between gauge-equivalent output-supported representatives. The single-shot derivation assumes metacheck consistency implies syndrome validity without requiring complete metachecks or otherwise establishing that implication.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 31 pages, 8 figures. Improved presentation and added new results