Lemma 6: Time Fault-Distance #
Statement #
The fault-distance for pure measurement and initialization errors (time-faults only, no space-faults) in the fault-tolerant gauging measurement procedure (Def 10) is $t_o - t_i = d$, the number of rounds in the deformed code phase.
Main Results #
timeFaultDistance: The time fault-distance (minimum weight of pure-time logical fault)gaussStringFault: The A_v measurement string of weight d (upper bound witness)gaussStringFault_weight_eq_d: The A_v string has weight dgaussStringFault_syndromeFree: The A_v string is syndrome-freetimeFaultDistance_le_d: d_time ≤ d (upper bound)timeFaultDistance_ge_d: d_time ≥ d (lower bound)timeFaultDistance_eq_d: d_time = d (main theorem)
Proof Outline #
Upper bound: A string of A_v measurement faults for a single vertex v at all d rounds of Phase 2 has weight d, is syndrome-free (repeated detectors see two flips that cancel), and flips σ (since d is odd).
Lower bound: Any pure time-fault that changes the logical outcome must flip σ. By the all-or-none constraint from repeated Gauss detectors, each vertex's A_v faults span either all d rounds or none. A sign flip requires an odd number of full A_v strings, each contributing d to weight.
Part I: Pure-Time Logical Fault Definition #
A pure-time spacetime fault: no space-faults, only time-faults.
Equations
- TimeFaultDistance.IsPureTimeFault proc F = F.isPureTime
Instances For
A pure-time gauging logical fault: pure-time AND a gauging logical fault.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The set of weights of pure-time gauging logical faults.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The time fault-distance: minimum weight of a pure-time gauging logical fault.
Equations
- TimeFaultDistance.timeFaultDistance proc detectors outcomePreserving = sInf (TimeFaultDistance.pureTimeLogicalFaultWeights proc detectors outcomePreserving)
Instances For
Part II: The A_v Measurement String (Upper Bound Witness) #
The A_v measurement string: time-faults on A_v at all d rounds of Phase 2.
Equations
- TimeFaultDistance.gaussMeasFaults proc v = Finset.image (fun (r : Fin proc.d) => { measurement := FTGaugingMeasurement.phase2 (DeformedCodeMeasurement.gaussLaw v r) }) Finset.univ
Instances For
The A_v measurement string as a spacetime fault.
Equations
- TimeFaultDistance.gaussStringFault proc v = { spaceFaults := ∅, timeFaults := TimeFaultDistance.gaussMeasFaults proc v }
Instances For
The A_v string is pure-time.
The image map for A_v faults is injective.
The A_v string has exactly d time-faults.
The A_v string has weight d.
Membership lemmas #
A time-fault is in the A_v string iff it's a Gauss A_v measurement at some round.
A Gauss A_v measurement at round r is in the A_v string.
A Gauss A_w measurement (w ≠ v) is not in the A_v string.
The A_v string flips the gauging sign #
The gaussSignFlip for the A_v string equals d (mod 2).
When d is odd, the A_v string flips the gauging sign.
Syndrome-freeness of the A_v string #
The A_v string is syndrome-free with respect to all detectors from Lemma 4.
Part III: Upper Bound #
The A_v string is a gauging logical fault when d is odd.
Upper bound: timeFaultDistance ≤ d when d is odd and V is nonempty.
Part IV: Lower Bound #
For a pure-time fault, weight = timeFaults.card.
Gauss fault count for vertex v: number of A_v measurement faults across all rounds.
Equations
- TimeFaultDistance.gaussFaultCount proc v F = {r : Fin proc.d | { measurement := FTGaugingMeasurement.phase2 (DeformedCodeMeasurement.gaussLaw v r) } ∈ F.timeFaults}.card
Instances For
The gaussSignFlip equals the sum of Gauss fault count parities.
Consecutive round constraint #
For a syndrome-free fault, consecutive Gauss A_v rounds have the same fault status.
All-or-none: for a syndrome-free fault, either all rounds of A_v are faulted or none are.
The Gauss fault count for each vertex is either 0 or d.
Gauss fault parity for each vertex is either 0 or d (mod 2).
For a syndrome-free fault that flips the sign with odd d: odd number of full A_v strings.
The total Gauss fault count ≤ weight for a pure-time fault.
A pure-time fault with a full A_v string has weight ≥ d.
Lower bound: Any pure-time logical fault has weight ≥ d (when d is odd).
Part V: Main Theorem #
Lower bound for time fault-distance: timeFaultDistance ≥ d.
Lemma 6 (Time Fault-Distance): timeFaultDistance = d.
Part V-b: Justification of the Outcome Predicate (Paper Step 5, second bullet) #
The paper's proof (Step 5) shows that for pure-time faults, PreservesGaugingSign is the
correct outcome predicate. Specifically: every syndrome-free pure-time fault that does NOT
flip σ decomposes into the spacetime stabilizer generators from Lemma 5 (B_p measurement
pairs, s̃_j measurement pairs, and boundary initialization/readout faults). Since these
generators preserve the gauging outcome by listedGenerator_isGaugingStabilizer (Lem 5),
they are trivial logical faults.
This section proves this equivalence, connecting our use of PreservesGaugingSign proc
to the paper's full argument.
Every syndrome-free pure-time fault that preserves the Gauss sign is a gauging stabilizer that decomposes into Lemma 5's listed generators. This is the mathematical content of Step 5 (second bullet) of the paper's proof: non-σ-flipping syndrome-free pure-time faults are products of spacetime stabilizers (B_p/s̃_j measurement pairs and boundary faults), hence trivial.
For syndrome-free pure-time faults, the dichotomy from Def 11 specializes to:
either the fault flips σ (logical fault), or it decomposes into Lemma 5's
stabilizer generators (trivial fault). This justifies using PreservesGaugingSign
as the outcomePreserving predicate in the time fault-distance definition.
Part VI: Corollaries #
Any pure-time fault of weight < d is a gauging stabilizer (when d is odd).
The Gauss fault count dichotomy (rephrased).
The time fault-distance equals the Phase 2 duration t_o - t_i.
Minimum-weight pure-time logical faults have exactly one full A_v string.