pith. sign in

arxiv: 2605.27262 · v1 · pith:3VGGYNLKnew · submitted 2026-05-26 · 🪐 quant-ph

Nonasymptotic bounds for quantum purity amplification

Thilo Scharnhorst , Jack Spilecki , John Wright This is my paper

Pith reviewed 2026-06-29 17:36 UTC · model grok-4.3

classification 🪐 quant-ph
keywords quantum purity amplificationnonasymptotic boundssample complexityfidelityprincipal eigenstateYoung diagramseigenvalue gaptensor powers
0
0 comments X

The pith

If a quantum state's largest eigenvalue strictly exceeds the second, O(k + (k/δ)·(1-p_d)/(p_d-p_{d-1})^2) copies suffice to prepare k copies of its principal eigenstate at fidelity 1-δ.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes nonasymptotic bounds for quantum purity amplification. Given n copies of a noisy state ρ with sorted eigenvalues p1 ≤ ⋯ ≤ pd and gap p_{d-1} < p_d, n = O(k + (k/δ) · (1-p_d)/(p_d-p_{d-1})^2) copies suffice to output a state with fidelity at least 1-δ to |v_d^{\otimes k}⟩. The bound holds for arbitrary spectra and is independent of dimension d. A sympathetic reader cares because earlier results gave only asymptotic guarantees as n tends to infinity, while this supplies explicit finite-sample requirements that match optimal scaling for depolarizing noise.

Core claim

The central claim is that if ρ's eigenvalues satisfy p1 ≤ ⋯ ≤ pd with p_{d-1} < p_d, then n = O(k + (k/δ) · (1-p_d)/(p_d-p_{d-1})^2) copies of ρ suffice to output a quantum state whose fidelity with |v_d^{\otimes k}⟩ is at least 1-δ. The guarantee is nonasymptotic, applies to any spectrum, and does not depend on d. The argument rests on the combinatorics of random Young diagrams.

What carries the argument

The combinatorics of random Young diagrams, which is used to bound the probability that a random diagram encodes the information needed to isolate the dominant eigenvector across tensor powers.

If this is right

  • The bound is independent of the Hilbert-space dimension d.
  • For depolarizing noise the finite-sample expression recovers the optimal asymptotic scaling.
  • The result applies to any eigenvalue spectrum provided only that the largest eigenvalue is strictly dominant.
  • The output state is guaranteed close in fidelity to the k-fold tensor power of the principal eigenvector.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • When the gap p_d - p_{d-1} shrinks, the quadratic dependence on its reciprocal suggests that sample requirements rise sharply near degeneracy.
  • The Young-diagram technique could be adapted to bound sample complexity for other symmetric tensor-power tasks such as spectrum estimation or state certification.
  • Numerical checks on small-dimensional systems with engineered gaps would directly test whether the O-notation hides only modest constants.

Load-bearing premise

The combinatorial analysis of random Young diagrams yields the claimed nonasymptotic sample bound under the strict eigenvalue gap p_{d-1} < p_d.

What would settle it

For a fixed two-qubit state with eigenvalues 0.7 and 0.3, compute the smallest n such that an explicit algorithm achieves fidelity 0.9 for k=2 and δ=0.1, then check whether that n lies within a small constant factor of the formula evaluated at those parameters.

Figures

Figures reproduced from arXiv: 2605.27262 by Jack Spilecki, John Wright, Thilo Scharnhorst.

Figure 1
Figure 1. Figure 1: From left to right: a Young diagram of shape [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Illustration of the overhang bi := µi − λi+1, shown for a fixed SSYT T of shape λ = (6, 3, 1) over the alphabet [3]. Removing the darker gray boxes, which contain the letter d = 3, leaves the sub-tableau T <d of shape µ = (5, 2, 0). The overhang bi counts the lighter gray boxes in row i: among the λi − λi+1 boxes in row i extending beyond row (i + 1), the overhang counts those that are not labeled d. In th… view at source ↗
read the original abstract

In quantum purity amplification, one is given $n$ copies of a noisy quantum state $\rho \in \mathbb{C}^{d \times d}$ and asked to prepare $k$ copies of its principal eigenstate $|v_d\rangle$. Several prior works have derived information-theoretically optimal algorithms for this problem, but the bounds they prove are only shown in the asymptotic regime as the number of samples $n$ tends to infinity. In this paper, we establish the following nonasymptotic guarantee: if $\rho$'s eigenvalues are sorted $p_1 \leq \cdots \leq p_d$ and $p_{d-1} < p_d$, then \begin{equation*} n = O\Big(k + \frac{k}{\delta} \cdot \frac{1-p_d}{(p_d-p_{d-1})^2}\Big) \end{equation*} copies suffice to output a state with fidelity at least $1-\delta$ with $|v_d^{\otimes k}\rangle$. Our bound holds for arbitrary spectra, and is independent of the dimension $d$. In the case of depolarizing noise, our finite-sample guarantee matches the optimal asymptotic scaling. Our proof is based on the combinatorics of random Young diagrams.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper claims a nonasymptotic guarantee for quantum purity amplification: given n copies of ρ ∈ ℂ^{d×d} with sorted eigenvalues p1 ≤ ⋯ ≤ pd and strict gap p_{d-1} < p_d, n = O(k + (k/δ)·(1−p_d)/(p_d−p_{d-1})^2) suffices to output a state with fidelity ≥1−δ to |v_d⟩^{\otimes k}. The bound is asserted to hold for arbitrary spectra and to be independent of dimension d; the proof relies on the combinatorics of random Young diagrams from the symmetric-group decomposition of ρ^{\otimes n}. In the depolarizing case the finite-sample bound matches the known optimal asymptotic scaling.

Significance. If the claimed dimension-independent bound is correct, the result would supply the first explicit nonasymptotic sample-complexity guarantee for this task, extending prior asymptotic optimality statements to finite n and k. The combinatorial Young-diagram technique, if it indeed cancels all d-dependent counting factors, could become a useful tool for other symmetric-subspace problems in quantum information.

major comments (2)
  1. [Abstract (displayed equation)] Abstract (displayed equation): the headline claim that the O(…) bound is independent of d is load-bearing. Standard counting via the hook-length formula or RSK correspondence for partitions with at most d rows produces d-dependent factors; the manuscript must exhibit an explicit tail-probability or union-bound estimate showing these factors cancel exactly. No such cancellation lemma is visible in the abstract, and the combinatorial argument therefore requires verification in the main proof.
  2. [Proof of the main theorem (combinatorial analysis of random Young diagrams)] Proof of the main theorem (combinatorial analysis of random Young diagrams): the probability that a random diagram fails to isolate the principal eigen-component must be bounded without residual d-dependent overhead. The argument should state the precise concentration inequality used and confirm that the union bound over the d-row partitions does not re-introduce dimension dependence.
minor comments (2)
  1. [Abstract] Notation for the principal eigenvector |v_d⟩ and its eigenvalue p_d should be introduced once and used consistently; the current abstract phrasing leaves the indexing convention implicit.
  2. [Introduction] The relation between the new finite-sample bound and the cited asymptotic works should be stated more quantitatively (e.g., by recovering the asymptotic scaling as δ→0 or k fixed).

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and for highlighting the need to make the dimension-independence argument fully explicit. We respond to each major comment below and will revise the manuscript accordingly.

read point-by-point responses

  1. Referee: [Abstract (displayed equation)] Abstract (displayed equation): the headline claim that the O(…) bound is independent of d is load-bearing. Standard counting via the hook-length formula or RSK correspondence for partitions with at most d rows produces d-dependent factors; the manuscript must exhibit an explicit tail-probability or union-bound estimate showing these factors cancel exactly. No such cancellation lemma is visible in the abstract, and the combinatorial argument therefore requires verification in the main proof.

    Authors: We agree that the abstract is too terse on this point. The dimension independence follows from an explicit tail bound in the proof of the main theorem (Section 3): after applying the hook-length formula to the probability of a random Young diagram, the d-dependent prefactors arising from the number of partitions with at most d rows are absorbed into the exponential decay term controlled by the eigenvalue gap (p_d − p_{d−1}). The resulting bound on the failure probability is O(exp(−c n (p_d − p_{d−1})^2 / (1−p_d))) with c independent of d. In the revision we will add a one-sentence pointer in the abstract to this cancellation and include a short dedicated paragraph (new Lemma 3.3) that isolates the union-bound estimate. revision: yes

  2. Referee: [Proof of the main theorem (combinatorial analysis of random Young diagrams)] Proof of the main theorem (combinatorial analysis of random Young diagrams): the probability that a random diagram fails to isolate the principal eigen-component must be bounded without residual d-dependent overhead. The argument should state the precise concentration inequality used and confirm that the union bound over the d-row partitions does not re-introduce dimension dependence.

    Authors: The proof already invokes the standard concentration result for the Plancherel measure on Young diagrams (the Vershik–Kerov–Logan–Shepp limit shape plus a large-deviation tail from [Biane 2001]). The union bound is taken over the set of diagrams with at most d rows; the cardinality of this set contributes at most d^{O(k)} but is multiplied by an exponentially small probability for any diagram whose first row is not sufficiently dominant. Because the gap condition makes the exponential decay linear in n and quadratic in the gap, the polynomial-in-d prefactor is dominated for the stated sample size. We will revise the proof to (i) name the concentration inequality explicitly and (ii) display the union-bound calculation as a separate displayed equation so that the cancellation is immediate. revision: yes

Circularity Check

0 steps flagged

No circularity; bound derived from independent combinatorial analysis of Young diagrams.

full rationale

The paper states that the nonasymptotic sample bound is obtained from the combinatorics of random Young diagrams in the decomposition of ρ^{\otimes n} under the symmetric group. No self-citation is used to justify the central claim, no parameters are fitted from data and then relabeled as predictions, and no ansatz or uniqueness result is imported from prior work by the same authors. The derivation is presented as a direct combinatorial argument that yields the d-independent guarantee for arbitrary spectra under the eigenvalue gap condition. This constitutes a self-contained mathematical derivation with no reduction of the output to the inputs by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the applicability of random Young diagram combinatorics to the quantum state amplification problem, which is invoked without further justification in the abstract.

axioms (1)
  • domain assumption Combinatorics of random Young diagrams yields the stated nonasymptotic sample bound for purity amplification
    Explicitly named as the basis of the proof in the abstract.

pith-pipeline@v0.9.1-grok · 5748 in / 1080 out tokens · 44868 ms · 2026-06-29T17:36:57.742817+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

6 extracted references · 5 canonical work pages · 3 internal anchors

  1. [1]

    Approximate factorization and concentration for characters of symmetric groups

    [Bia01] Philippe Biane. Approximate factorization and concentration for characters of symmetric groups. International Mathematics Research Notices, 2001(4):179–192,

    2001

  2. [2]

    Streaming quantum state purification for general mixed states

    [GLL+25] Daniel Grier, Debbie Leung, Zhi Li, Hakop Pashayan, and Luke Schaeffer. Streaming quantum state purification for general mixed states. Technical report, arXiv:2503.22644,

    work page arXiv

  3. [3]

    Optimal quantum purity amplifica- tion

    [LFIC24] Zhaoyi Li, Honghao Fu, Takuya Isogawa, and Isaac Chuang. Optimal quantum purity amplifica- tion. Technical report, arXiv:2409.18167,

    work page arXiv

  4. [4]

    An Exponential Sample-Complexity Advantage for Coherent Quantum Inference

    [LTHC26a] Zhaoyi Li, Elias Theil, Aram Harrow, and Isaac Chuang. An exponential sample-complexity advantage for coherent quantum inference. Technical report, arXiv:2605.21457,

    work page internal anchor Pith review Pith/arXiv arXiv

  5. [5]

    Quantum Purity Amplification for Arbitrary Eigenstates and Multiple Outputs

    [LTHC26b] Zhaoyi Li, Elias Theil, Aram Harrow, and Isaac Chuang. Quantum purity amplification for arbitrary eigenstates and multiple outputs. Technical report, arXiv:2605.21570,

    work page internal anchor Pith review Pith/arXiv arXiv

  6. [6]

    Kerov's central limit theorem for Schur-Weyl measures of parameter 1/2

    [M´ el10a] Pierre-Lo¨ ıc M´ eliot. Kerov’s central limit theorem for Schur-Weyl measures of parameter 1/2. Technical report, arXiv:1009.4034,

    work page internal anchor Pith review Pith/arXiv arXiv