Skip to content

About

Exact quantum-memory limits for calibrating a randomized Pauli detector, with attaining protocols and adaptive estimation bounds.

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Latest commit

 

History

9 Commits

Folders and files

Repository files navigation

Quantum Detector Memory

How much quantum memory is needed to calibrate a detector that destroys the probe and returns only a classical record?

For a known detector family with one unknown contrast, this project gives an exact retained-memory information frontier and a matching global estimation law. The detector privately selects a fresh random Pauli measurement on each use and reveals the setting only after consuming the probe.

Let $d=2^n$ be the probe dimension and $r\leq d$ the dimension of all retained quantum systems during a call. The effects are

$$E_{P,s}(\theta)=\frac{I+s\theta P}{2(d^2-1)}, \qquad P\ne I,\quad s\in\{-1,+1\},\quad |\theta|\lt1,$$

where $P$ ranges over Hermitian $n$-qubit Paulis. Define $c_r=(dr-1)/(d^2-1)$. The exact optimal one-use Fisher information is

$$F_r^\star(\theta)=\frac{c_r}{1-\theta^2}.$$

For each integer $r$, one fixed input and a setting-conditioned reference measurement attain this simultaneously for every interior contrast. The resulting experiment is an observed erasure with probability $1-c_r$, followed on success by a Bernoulli sign of mean $\theta$.

Principal results

Question Result Read
How much one-use information can a reference retain? Exact Fisher coefficient $c_r$ for all $r$ and interior contrasts Fisher frontier
Is the frontier physically attainable? Uniform entanglement over a computational subset; $k$ Bell pairs suffice when $r=2^k$ Construction and resources
Can coherent adaptive reuse improve distinguishability from zero? Exact maximum relative entropy $Nc_rh(\theta)$ after $N$ calls Entropy and adaptive calibration
What is the global estimation cost? Minimax mean-square error within a factor 32 of $\min\{a^2,1/(Nc_r)\}$ on $[-a,a]$ Finite-sample guarantees
Can one scalar detector parameter benefit from an ancilla? A six-outcome qubit detector has an exact factor-three Fisher advantage Qubit counterexample

Here $h(\theta)=[(1+\theta)\log(1+\theta)+(1-\theta)\log(1-\theta)]/2$ and logarithms are natural. The adaptive converses allow arbitrary parameter-independent controls, free classical history, mixed states, intermediate measurements and reference reuse.

With $k$ retained qubits, root-mean-square accuracy $\epsilon$ on a fixed interval containing zero requires and admits

$$N_\epsilon =\Theta\!\left(\frac{2^{2n}-1}{(2^{n+k}-1)\epsilon^2}\right) =\Theta\!\left(\frac{2^{n-k}}{\epsilon^2}\right)$$

in the nontrivial small-error regime. Uniform high-confidence estimation adds a factor $\log(1/\delta)$; explicit constants and ranges are provided.

Read the science

Start with the model, assumptions and exact claims, then the attaining protocol. Specialists can go directly to the Fisher proof, relative-entropy proof, or adaptive minimax argument.

The interpretation explains why disclosure timing matters and compares the result with depolarizing-channel estimation. The related-work account and structured bibliography credit the inherited local Schmidt form, ancillary calibration, Pauli-learning memory tradeoffs and replacer-channel amortization. The LLM reading guide maps questions to their authoritative files.

The new claim is the solved destructive-detector calibration model and its consequences. Generic ancillary advantage, the zero-contrast rank numerator and exponential memory tradeoffs have established precedents.

Learn the project

The single educational anchor is John Watrous, The Theory of Quantum Information (2018). The learning entrypoint connects its foundations to this detector model without assigning additional textbooks.

Route What it provides
Two-qubit worked experiment Probe preparation, all fifteen settings, observed erasures, the Y-sign correction and a complete estimator
Watrous-to-project bridge Verified source sections, notation and log conventions, and the locally supplied Fisher, entropy and statistical arguments
Thirty-milestone map Twelve background milestones and eighteen narrative milestones, with understanding targets, dependencies and canonical proof pointers

The map is learner independent. It separates local Fisher information, global relative-entropy bounds and minimax rates; the scientific pages above remain the authoritative proofs.

Reproduce the finite checks

From the repository root:

python3 -m pip install -r verification/requirements.txt
python3 verification/run_all.py

The verification guide documents the actual matrix checks, deterministic cases, expected tolerances and scope. These corroborate finite-dimensional calculations; the universal adaptive and minimax statements rest on the written proofs.

Scope that matters

The setting is disclosed after preparation. Every accessible retained coherent register, future entangled probe and purification counts in $r$. Preparation and reference readout are trusted; detector calls are independent. Query budgets count failures and are worst-case budgets. The explicit powers-of-two implementation uses local gates and measurements; efficient arbitrary-$r$ encoding is not asserted.

The exact adaptive Fisher optimum is established at zero contrast. Exact relative-entropy optimality does not imply exact optimality for every finite-sample testing loss, and the global minimax theorem concerns intervals containing zero. Returning the quantum probe or revealing the setting before preparation changes the resource model.

Purpose and contact

This repository serves as a record of the work and a guide for the author’s self-directed learning. For discussion or potential collaboration, please contact Ruge Lin at gogoko699@gmail.com.

About

Exact quantum-memory limits for calibrating a randomized Pauli detector, with attaining protocols and adaptive estimation bounds.

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages