Tight Lower Bounds for State Tomography with Limited Entanglement
Ufuk Keskin, Jason Luo, Mahbod Majid, Matthew Radzihovsky
quant-ph · cs.IT · cs.LG
Quick verdict
Claims tight k-copy tomography lower bound max{d³/(√kε²), d²/ε²} for all k; rigorous, complete proof; hinges on two lemmas cited from CLL24 full version.
Summary
This paper resolves the copy complexity of quantum state tomography when each measurement acts on at most k fresh copies and no quantum memory is retained between blocks. The authors prove a lower bound n ≳ max{d³/(√k ε²), d²/ε²} for every k and all sufficiently small ε, matching the upper bound of Pelecanos–Spilecki–Wright [PSW26] and thereby settling, up to absolute constants, the tradeoff left partially open by Chen–Li–Liu [CLL24], whose lower bound required k ≤ (1/ε)^γ. The bound covers fully adaptive protocols, random and nonuniform block sizes, and expected copy budgets, and already applies in a small neighborhood of any state with smallest eigenvalue of order 1/d, even when that center is known. The key new ingredient is a uniform Fisher information bound for a single t-copy measurement block, tr J_M(ρ) ≤ min{2t^{3/2}/λmin(ρ)², t(d²−1)/λmin(ρ)} (Theorem 8.1), proved via a change of variables from Zhu–Hayashi and two Schur–Weyl estimates taken from the CLL24 full version, combined with a centering step at the ensemble mean tρ that cancels the t²tr(ρ²) term and eliminates the perturbative expansion behind CLL24's restriction. Limitations: results hold only for ε ≤ ε0 (an implicit constant), two load-bearing lemmas are cited rather than re-proven, and no explicit constants are given.
Review complete
Assessment at a glance
4
risk flags identified
● high 0
● medium 1
● low 3
4 strengths · 4 weaknesses · 5 author questions
AI text signal = Binoculars zero-shot detector score over 20 sampled passages (Fast-DetectGPT co-reported) — absence does not prove human authorship, and a flag does not imply misconduct.
Levels reflect verification complexity — not a quality score. Evolve never rates your paper.
Key claims under review
- Any protocol whose measurements act on at most k copies needs n ≳ max{d³/(√k ε²), d²/ε²} copies to learn a d-dimensional state to trace distance ε, for every k and all sufficiently small ε (Theorem 1.3).
- Every POVM on t copies of a full-rank state obeys the uniform Fisher information bound tr J_M(ρ) ≤ min{2t^{3/2}/λmin(ρ)², t(d²−1)/λmin(ρ)}, with the t^{3/2} branch being the new contribution (Theorem 8.1).
- The lower bound already holds locally, in an operator-norm ball of radius Lε/d around any state ρ⋆ ⪰ Id/(2d), even when the protocol knows ρ⋆ (Corollary 1.4).
- Combined with the [PSW26] upper bound, the minimax copy complexity with k-copy measurements is Θ(max{d³/(√k ε²), d²/ε²}) for every k at small ε (Corollary 1.6).
- The bound survives randomized stopping rules when the copy count is bounded in expectation over all input states (Corollary 1.5).
Strengths
Fully resolves a recently opened tradeoff
Theorem 1.3 matches the [PSW26] upper bound for every k with no relation between k and ε, upgrading CLL24's restricted lower bound (k ≤ (1/ε)^γ) to the Θ-characterization of Corollary 1.6, n⋆(d,k,ε) ≍ max{d³/(√k ε²), d²/ε²}.
Elegant new core lemma with a clean cancellation
Theorem 8.1's t^{3/2}/λmin(ρ)² bound works by centering the unfolded matrix at its POVM-forced mean tρ (Eq. (28)), cancelling the t²tr(ρ²) term of the Schur–Weyl moment bound against t(t−1)tr(ρ²), leaving 2t^{3/2} (Eq. (14), Lemma 8.2) — removing CLL24's perturbative expansion and its k–ε restriction.
Unusually complete corollary structure
Above-norm rigor hygiene
Weaknesses
Dependence on unproven external lemmas
Small-accuracy regime only
No explicit main constants
Scope delimited to λmin ~ 1/d instances
Full Review Report
$1.99 · one-timeUnlock the complete preliminary analysis Evolve produced for this paper.
Report preview
Scope & Claims
The paper settles the copy complexity of quantum state tomography in the limited-entanglement model (Definition 1.1): each measurement acts on at most k fresh copies, only classical information persists between blocks, and every POVM may adapt to prior outcomes. The central claim (Theorem 1.3) is that n ≳ max{d³/(√k ε²), d²/ε²} copies are necessary for every k and all ε ≤ ε0, matching the [PSW26] upper bound and yielding minimax optimality up to absolute constants (Corollary 1.6). Two refinements strengthen the result: Corollary 1.4 shows the bound already binds in an operator-norm ball of radius Lε/d around any state ρ⋆ ⪰ Id/(2d), even when ρ⋆ is known to the protocol; Corollary 1.5 extends from almost-sure to expected copy budgets via a 6n-truncation and Markov's inequality. The framing — removing CLL24's restriction k ≤ (1/ε)^γ — is accurate and fairly credited in Section 3.
The remaining sections cover Methodology Assessment, Statistical Evidence, Results & Reproducibility, Limitations & Risks, and Verdict & Recommendations.
- All 4 strengths & 4 weaknesses, analyzed in detail
- 4 risk flags with severity, section & reviewer questions
- Methods & statistical evidence assessment
- 5 questions a peer reviewer would ask
- Full six-dimension evaluation report & recommendations
One-time purchase · this paper only
One-time purchase, this paper only. AI-generated preliminary analysis — not peer review.
Frequently asked questions
What is the copy complexity of quantum state tomography when each measurement can use at most k copies?
For d-dimensional states at trace distance ε (with ε sufficiently small), this paper proves the matching lower bound max{d³/(√k ε²), d²/ε²}. Together with the known upper bound of Pelecanos–Spilecki–Wright, the answer is determined up to constant factors: jointly measuring √k copies at a time saves a factor √k until k reaches d², beyond which the unrestricted d²/ε² rate applies.
How does this paper improve on the Chen–Li–Liu (CLL24) result?
CLL24 proved the d³/(√k ε²) lower bound only when k ≤ (1/ε)^γ for a small constant γ, because their proof used a perturbative expansion of the state. This paper removes that restriction entirely, proving the lower bound for every k, by establishing a uniform Fisher information bound per measurement block that avoids perturbation.
Does adaptivity help a protocol beat this lower bound?
No. The lower bound applies to fully adaptive protocols that choose each block's POVM based on all previous classical outcomes, allow random or nonuniform block sizes, may stop at a random time, and even remain valid when only the expected total number of copies is bounded.
Was AI used to produce this proof?
Yes, and it is disclosed in the paper: the authors state that the main ideas and proof strategies were developed with assistance from OpenAI Codex (GPT-5.6 Sol model), that they studied, refined, simplified and verified the material, and that they take responsibility for every claim, proof, and citation.
About this review
This is an AI-generated preliminary review by Evolve.Pub, produced from the openly available text of the paper. It is not peer review and does not represent the authors' views. The canonical record of the paper is at the publisher link above.
- PDF extraction flattens inline math (superscripts, norms, subscripts such as C6.1); formulas were read from the flattened text and appear internally consistent, but symbol-level fidelity is reconstructed from context.
- Citation keys render with stray spaces (e.g., '[PSW26 ]') throughout — a bibliography extraction artifact, not missing content; the reference list is present and complete through [ZH18].
Go deeper on this paper
Wrote a paper? Stress-test it before reviewers do.
Upload your manuscript for the same structured scrutiny — reasoning, statistics, evidence, figures, and reproducibility.