A uniform compiler for measuring a signed output register
ProvedInfoLossLAR.WorstCase.RegevSearch.compile_signed_register_readoutcomplexityquantum-computing
Let a polynomial-time program assign to a binary input a quantum circuit on wires, an offset , and an output shape of signed words of bits each. The assignment uses the native Hadamard, phase, , bit-flip, and controlled bit-flip instructions. There is a uniform quantum circuit family with a polynomial-time description generator which implements the readout on every correctly encoded circuit with . For any finite set of integer vectors, its measured probability of decoding a member of is
With probability one the output contains the canonical vector encoding within the reserved output prefix. Both the dimension and the register widths may depend on the input.
Preamble
import Definitions.Def_InfoLossLAR_RegevSearch_UniformFourierRow set_option autoImplicit false set_option backward.isDefEq.respectTransparency false set_option maxHeartbeats 1500000 open scoped Classical BigOperators
Formal statement
namespace InfoLossLAR.WorstCase.RegevSearch
open QAlgorithms QAlgorithms.Complexity GaussianIteration AdaptiveStages
open OAI.MetricKMedian.Binary OAI.MetricKMedian.Binary.Automatic
theorem compile_signed_register_readout
(f : Str → (List Unit × List Unit) × ((List Unit × ℕ) × List (ℕ × ℕ × ℕ))) (hf : Poly f) :
∃ (G : MFamily) (desc : Str → Str), Poly desc ∧ (∀ n,desc (unary n)=G.encAt n) ∧ IsMQPT G ∧
∀ (w : Str) (v : (List Unit × List Unit) × ((List Unit × ℕ) × List (ℕ × ℕ × ℕ))), f w=v →
∀ (s : List Unit × List Unit), v.1=s → ∀ off : ℕ, v.2.1.2=off →
∀ c : Layered v.2.1.1.length, v.2.2=c.flatten.map (fun g => match g with
| .h i => (0,i.val,0) | .s i => (1,i.val,0) | .t i => (2,i.val,0)
| .x i => (3,i.val,0) | .cnot i j _ => (4,i.val,j.val)) →
off+s.1.length*s.2.length≤v.2.1.1.length →
(∀ S, (SignedSlot.slot s).length≤S → outputProb G w (SignedSlot.canon s.1.length S)=1) ∧
∀ T : Finset (Vec s.1.length),
finiteMass ((measuredWord G w).map (SignedSlot.decodeWord s.1.length)) T =
∑ x, if ∃ z∈T, encVec encI z <+: GridReadout.Axis.encodeSlices (s.1,s.2,(List.ofFn x).drop off)
then ‖runLayered c (fun y => if y=(fun _=>false) then 1 else 0) x‖^2 else 0 := by sorry
end InfoLossLAR.WorstCase.RegevSearchSource
InfoLossLAR formalization of the signed-register readout used in the shared-pool Fourier sampler. https://prove2.me/theorems/d106aba4-0ad7-4239-9c6b-4a9dc8c9ebc4