Dimension-Free Polylogarithmic Quantum Shadow Tomography from Sequential Pretty-Good Measurements

Fernando Granha Jeronimo; Lenny Liu; Qizhao Huang

Dimension-Free Polylogarithmic Quantum Shadow Tomography from Sequential Pretty-Good Measurements

T0 review · 0 major / 7 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read A sequential pretty-good measurement protocol achieves dimension-free polylogarithmic shadow tomography.

desk verdict Genuine resolution of Aaronson's dimension-free polylog shadow tomography question; the core posterior-contraction proof holds up on inspection, with a fixable misattribution of the prior best bound. read the letter →

arxiv 2608.06345 v1 pith:C2CRI2LP submitted 2026-08-06 quant-ph

Fernando Granha Jeronimo , Qizhao Huang , Lenny Liu This is my paper

classification quant-ph MSC 81P6868Q12 PACS 03.67.-a

keywords shadowtomographysamplecomplexitypretty-goodmeasurementsequentialposteriorconcentrationminimaxdimension-freeboundquantumstatediscrimination

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

Shadow tomography asks how many copies of an unknown quantum state are needed to estimate the expectation values of many known observables. This paper establishes that the number of copies can be polylogarithmic in the number of observables and completely independent of the Hilbert-space dimension, namely $O(\varepsilon^{-2} (\log(M/\delta))^4 / (\log\log(M/\delta))^3)$ copies. That resolves the open question from the original shadow-tomography paper, and it improves exponentially on the previous best dimension-free bound, which scaled roughly as $\sqrt{M}$. The proof solves a finite-ensemble Bayesian version via sequential pretty-good measurements, then removes the prior and the finite-state restriction through a minimax argument and a trace-distance net.

What carries the argument

Three pieces carry the argument. The first is the completed pretty-good measurement on blocks of $n$ copies, whose operators are $\bar\rho^{-1/2} q_x \rho_x^{\otimes n} \bar\rho^{-1/2}$ plus a support term; the sequential variant updates the prior by the posterior after each block. The second is the factor-two quadratic comparison, which states that for any real labels attached to the ensemble, the mean-square error of decoding the pretty-good measurement outcome is at most twice the minimum mean-square error of any scalar measurement. The load-bearing third piece is the posterior concentration deficit $R_s(q;E)$, the posterior mass left outside the best interval of radius $s$ for $\mathrm{Tr}(E\rho)$. Theorem 4.1 shows that one block of $n$ copies contracts this deficit to $\frac{2}{n\gamma^2}R_s(q;E) + \frac{8}{n\gamma^2}e^{-n\gamma^2/2}$ when the allowed interval is widened by $\gamma$. Iterating that contraction gives polylogarithmic concentration; a common finite decoder, a minimax theorem, and a trace-distance net then lift the Bayesian guarantee to the worst-case, dimension-free theorem.

What would settle it

Numerically evaluate the exact one-block contraction: take a two- or three-state ensemble, choose a prior $q$, an observable $E$, tolerances $s,\gamma>0$, and block size $n$, compute the conditional pretty-good-measurement posterior distributions, and check whether the expected value of $R_{s+\gamma}(q_Y;E)$ ever exceeds $\min\{R_s(q;E), \frac{2}{u}R_s(q;E)+\frac{8}{u}e^{-u/2}\}$ with $u=n\gamma^2$. A single violation would disprove Theorem 4.1 and remove the foundation of the copy bound.

Watch

Extended reading notes

Core claim

The paper's central claim is that a single collective measurement strategy solves shadow tomography with sample complexity $T \le C \varepsilon^{-2} L^4/(\log L)^3$, where $L = \log(16M/\delta)$, for every finite-dimensional Hilbert space, every list of observables $0 \preceq E_j \preceq I$, every unknown state, and all accuracy and failure parameters. The strategy depends on the observable list but not on the unknown state. The construction first solves a Bayesian version: for a finite ensemble of candidate states, an $r$-round sequential pretty-good measurement replaces the prior by the posterior after each block of fresh copies, and a refined tail analysis shows all expectation estimates concentrate simultaneously. A minimax argument converts this finite-prior guarantee into one measurement uniform over any finite family of states, and a trace-distance net extends the same measurement to all states without introducing any dimension factor into the copy count.

Load-bearing premise

The polylogarithmic rate depends on the one-block posterior contraction bound of Theorem 4.1, which asserts that after one block of $n$ fresh copies the posterior mass outside the best radius-$s$ interval shrinks by the factor $2/(n\gamma^2)$ plus an exponentially small term; if that contraction is weaker in $n$ or $\gamma$, the dimension-free polylogarithmic conclusion no longer follows.

Editorial extensions

If this is right

For every finite-dimensional Hilbert space, every list of $M$ observables, and every unknown state, the protocol achieves accuracy $\varepsilon$ and failure probability $\delta$ using $O(\varepsilon^{-2}(\log(M/\delta))^4/(\log\log(M/\delta))^3)$ copies, with no dependence on dimension.
This copy count is polylogarithmic in $M$ and settles the original dimension-free question raised when shadow tomography was introduced.
It improves the previous best dimension-independent bound, which grew like $\sqrt{M}\log M/\varepsilon^2$, to a quantity that is exponentially smaller in $M$.
The same measurement works for all states simultaneously, not only for states drawn from a known prior distribution.
The result is information-theoretic: the paper does not claim an efficient implementation of the measurement or an efficient classical decoder.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Tightening the one-block contraction in Theorem 4.1 may close the gap between the $\log^4/(\log\log)^3$ rate and the lower-bound order, since the analysis is not shown to be optimal in the logarithmic factors.
The same sequential posterior-refocused pretty-good measurement could plausibly be adapted to online shadow tomography, where observables arrive one at a time and a protocol must answer before seeing the next; the paper does not analyze that setting.
The factor-two comparison with signed-power and residual labels is a general estimation tool and may apply to estimating nonlinear functionals of an ensemble, not only linear expectation values.
A concrete stress test is to simulate small ensembles and check the one-block posterior contraction inequality directly, isolating the step on which the whole polylogarithmic rate rests.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

First thing to know: this paper appears to settle Aaronson's dimension-free polylog shadow-tomography question. The proof is intricate but the main mechanism is coherent, and I did not find a load-bearing error. The stress-test note agrees, and I concur.

What is genuinely new is the sequential pretty-good measurement that re-centers the prior on the posterior after each block, plus the 'posterior concentration deficit' analysis. Theorem 4.1 is the heart: after one block of n copies, the posterior mass outside a widened interval contracts by 2/u plus an exponentially small term. That is what makes polylog work. The factor-two quadratic comparison (Theorem 2.4) is proved in the paper rather than imported; the moment induction in Theorem 3.3 checks out; and the minimax-plus-net lift in Section 5 is standard and correctly applied. The final copy count O(eps^{-2} L^4/(log L)^3) follows from the stated parameters.

The soft spots are real but not fatal. The abstract and Section 1.1 credit the prior best dimension-independent bound to Sinha [Sin25] with O(sqrt(M log M)/eps^2), while the paper's own related-work section cites [COPW26] with the strictly better O(sqrt(M)/eps^2). So the 'exponential improvement' is measured against the wrong baseline. The main theorem does not depend on that comparison, but the framing is inaccurate and should be corrected. Also, the proof is long and not machine-checked; I did not verify every line, but the parts I checked are internally consistent, and the stress-test pass on Theorem 4.1 is reassuring. The background assumptions are standard: Sion, Hoeffding, binomial moments, and the MLW square-root measurement comparison. No circularity, no fitted parameters.

Who is this for? Anyone working on quantum state tomography, quantum hypothesis testing, or complexity of quantum measurements. If the result is correct—and I think it is likely correct—it removes a major open problem. It deserves a serious referee. I would recommend sending it to peer review, and asking the authors to fix the prior-work attribution before it appears.

Referee Report

0 major / 7 minor

Summary. The paper claims a dimension-free polylogarithmic sample complexity for quantum shadow tomography: for every finite-dimensional Hilbert space, every list of M observables, accuracy epsilon, and failure delta, there is a strategy using T = O(epsilon^{-2} L^4/(log L)^3) copies with L = log(16M/delta). The protocol is a sequential pretty-good measurement that repeatedly updates a Bayesian prior over a finite ensemble. The proof combines a factor-two quadratic comparison theorem, a higher-moment bound, a posterior contraction estimate, and a minimax plus trace-distance-net argument to lift the finite-prior guarantee to a worst-case guarantee over all states.

Significance. If correct, this resolves Aaronson's open question and improves the prior dimension-independent bound of O(sqrt(M)/epsilon^2) to a polylogarithmic-in-M rate with no dimension dependence. The paper is self-contained in the key step: Theorem 2.4 proves the factor-two PGM comparison rather than importing it as a black box, and Theorem 4.1's posterior contraction is the load-bearing step. I checked the contraction calculation, including the Hoeffding tail-integral and the subsequent iteration in Theorem 4.3, and found it internally consistent. The minimax uniformization (Lemma 5.1, Proposition 5.2) and the trace-distance-net lift (Theorem 5.4) are standard and correctly applied. There are no fitted parameters, no circularities, and the main theorem would constitute a major advance in quantum shadow tomography.

minor comments (7)

[Abstract and Section 1.1] The abstract and Section 1.1 describe the prior best dimension-independent bound as O(sqrt(M) log M / epsilon^2) due to Sinha [Sin25], but Section 1.2 cites [COPW26] with the better bound O(sqrt(M)/epsilon^2). Please correct the attribution so that the claimed exponential improvement is measured against the actual prior best.
[Section 5.3] The final paragraph should be expanded into an explicit proof of Theorem 1.2. In particular, specify the choices epsilon_inner = epsilon/2, beta = delta/2, and s = min{epsilon/2, delta/(2T)} before applying Proposition 5.2 and Theorem 5.4; as written, the parameter selection is only sketched.
[Section 3.2] The proof of Theorem 3.3 refers to 'Theorem 3.1' and 'Theorem 3.2' when the intended statements are Lemma 3.1 and Lemma 3.2; please update the cross-references.
[Lemma 3.1] The proof of Lemma 3.1 uses u and v both for the original variables and for their shifts by a; setting x = u - a and y = v - a at the start would remove the ambiguity.
[Section 1.3 and Section 3.3] The proof roadmap mentions a 'matching copy-exponent lower bound,' but Section 3.3 proves a lower bound on the moment constants rather than on the copy exponent; please rephrase to avoid overstating what is shown.
[Corollary 3.5] The proof writes 'Take r = sqrt(log(eM/delta))' but r must be an integer; please state that r is rounded to a suitable integer and that the stated bound holds after rounding.
[Abstract] There are several typographical errors, including 'quantum shadow tomographyis' and 'priori distribution'; a careful proofreading pass is needed.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the load-bearing posterior-contraction theorem is proved from the factor-two PGM comparison and standard tail bounds, not assumed.

full rationale

All load-bearing steps in the derivation are proved inside the paper or use standard external theorems in a non-question-begging way. The factor-two PGM comparison (Theorem 2.4) is not taken as a black box: it is proved in Section 2.3 from Proposition 2.3, which is itself proved. The one-block posterior contraction (Theorem 4.1), which carries the polylogarithmic rate, is derived by feeding signed residual labels to this proved comparison and bounding the hypothetical {E,I-E}^{⊗n} experiment with Hoeffding tail integrals; it is not defined as its own conclusion. The minimax lift (Lemma 5.1 and Proposition 5.2) uses Sion's theorem for a zero-sum game whose payoff is the failure probability, and the trace-distance net (Lemma 5.3, Theorem 5.4) is a standard Lipschitz extension. There are no fitted parameters renamed as predictions, and the copy-count bound is a theorem rather than a fit. The paper contains no load-bearing self-citations: all cited prior works are external, and the authors' own prior work is not invoked to exclude alternatives. The only notable flaw is an attribution inconsistency: the abstract names Sinha [Sin25] as the prior best, while Section 1.2 cites the better O(sqrt{M}/eps^2) bound of [COPW26]; that affects historical accuracy, not circularity. Score 0.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The central claim rests on standard external theorems (Sion's minimax, Hoeffding, trace-distance bounds) plus the factor-two PGM comparison, which the paper proves in Section 2.3. No data fitting, no ad hoc constants, and no new physical entities are introduced.

assumptions (6)

standard math Sion's minimax theorem for convex compact strategy spaces
Used in Lemma 5.1 to exchange min over POVMs and max over prior distributions.
standard math Hoeffding's inequality for bounded random variables
Used in Theorem 4.1 and Appendix A to bound binomial tail probabilities.
standard math Optimal scalar quantum estimator via Sylvester equation (Personick)
Proved as Proposition 2.3; forms the basis of the factor-two PGM comparison.
standard math Factor-two quadratic comparison for the pretty-good measurement (Theorem 2.4)
Specialization of [MLW25] proved in Section 2.3; central engine for all loss bounds.
standard math Trace-distance stability of expectation values and tensor powers (Lemma 5.3)
Used to extend guarantees from the trace-distance net to all states.
domain assumption Finite-dimensional compactness of the state space and existence of trace-distance nets
Assumes dim H finite; the theorem is stated for finite-dimensional Hilbert spaces.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Dimension-Free Polylogarithmic Quantum Shadow Tomography from Sequential Pretty-Good Measurements." pith.science (2026). https://pith.science/paper/C2CRI2LP

@misc{pith2026260806345,
  author       = {Pith},
  title        = {Pith review of: Dimension-Free Polylogarithmic Quantum Shadow Tomography from Sequential Pretty-Good Measurements},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/C2CRI2LP}},
  note         = {Machine review of arXiv:2608.06345}
}

abstract

\textit{Shadow tomography} is a fundamental problem in quantum information theory. Given multiple copies of an unknown $d$-dimensional quantum state $\rho$ and a known collection of observables ${E_1,\ldots,E_m}$, the goal is to estimate all expectation values $\{\Tr(\rho E_i)\}_{i=1}^m$ to additive accuracy $\varepsilon$ with probability at least $1-\delta$. An elusive open question from the seminal shadow tomography work of Aaronson (STOC'18) is whether this task admits a dimension-independent sample complexity with only polylogarithmic dependence on $m$, as suggested by the best-known lower bounds. In this work, we give a quantum protocol for shadow tomography with sample complexity \[ O\left( \frac{1}{\varepsilon^2} \frac{(\log (m/\delta))^4} {(\log\log (m/\delta))^3} \right), \] which is polylogarithmic in the number of observables and independent of the dimension of the unknown state thereby answering Aaronson's original question while also providing an exponential improvement in the prior best dimension independent sample complexity of shadow tomography from Sinha (STOC'25). Our approach first reduces the general shadow-tomography problem to a finite-ensemble estimation problem via a minimax argument. We then develop an observable-independent protocol that repeatedly applies the pretty-good measurement and updates the priori distribution over the finite ensemble according to the measurement outcomes. A refined tail analysis of the resulting estimation error yields simultaneous accuracy guarantees for all observables.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

44 extracted references · 33 canonical work pages

[1]

Proceedings of the 50th Annual ACM SIGACT Symposium on Theory of Computing , series =

Scott Aaronson , title =. Proceedings of the 50th Annual ACM SIGACT Symposium on Theory of Computing , series =. 2018 , doi =. 1711.01053 , archiveprefix =

arXiv 2018
[2]

Improved quantum data analysis

Costin B. Improved Quantum Data Analysis , journal =. 2024 , doi =. 2011.10908 , archiveprefix =

work page Pith review arXiv 2024
[3]

Quantum Event Learning and Gentle Random Measurements

Adam Bene Watts and John Bostanci , title =. 15th Innovations in Theoretical Computer Science Conference (ITCS 2024) , series =. 2024 , doi =. 2210.09155 , archiveprefix =

work page Pith review arXiv 2024
[4]

Dimension Independent and Computationally Efficient Shadow Tomography

Pulkit Sinha , title =. Proceedings of the 57th Annual ACM Symposium on Theory of Computing , series =. 2025 , doi =. 2411.01420 , archiveprefix =

work page Pith review arXiv 2025
[5]

arXiv preprint arXiv:2407.13874 , year =

Sitan Chen and Jerry Li and Allen Liu , title =. arXiv preprint arXiv:2407.13874 , year =. doi:10.48550/arXiv.2407.13874 , eprint =

arXiv doi:10.48550/arxiv.2407.13874
[6]

Proceedings of Thirty Ninth Conference on Learning Theory , series =

Senrui Chen and Weiyuan Gong and Sisi Zhou , title =. Proceedings of Thirty Ninth Conference on Learning Theory , series =. 2026 , url =. 2602.04952 , archiveprefix =

arXiv 2026
[7]

Gentle Measurement of Quantum States and Differential Privacy

Scott Aaronson and Guy N. Rothblum , title =. Proceedings of the 51st Annual ACM SIGACT Symposium on Theory of Computing , series =. 2019 , doi =. 1904.08747 , archiveprefix =

work page Pith review arXiv 2019
[8]

Sequential measurements, disturbance and property testing

Aram W. Harrow and Cedric Yen-Yu Lin and Ashley Montanaro , title =. Proceedings of the Twenty-Eighth Annual ACM-SIAM Symposium on Discrete Algorithms , pages =. 2017 , doi =. 1607.03236 , archiveprefix =

work page Pith review arXiv 2017

Show all 44 references

[10]

V. P. Belavkin , title =. Stochastics , volume =. 1975 , doi =

1975
[11]

Wootters , title =

Paul Hausladen and William K. Wootters , title =. Journal of Modern Optics , volume =. 1994 , doi =

1994
[12]

Journal of Mathematical Physics , volume =

Howard Barnum and Emanuel Knill , title =. Journal of Mathematical Physics , volume =. 2002 , doi =. quant-ph/0004088 , archiveprefix =

2002
[13]

Mishra and Ludovico Lami and Mark M

Hemant K. Mishra and Ludovico Lami and Mark M. Wilde , title =. arXiv preprint arXiv:2505.20447 , year =. doi:10.48550/arXiv.2505.20447 , eprint =

doi arXiv
[14]

Personick , title =

Stewart D. Personick , title =. IEEE Transactions on Information Theory , volume =. 1971 , doi =

1971
[15]

Helstrom , title =

Carl W. Helstrom , title =. 1976 , isbn =

1976
[16]

Holevo , title =

Alexander S. Holevo , title =. Journal of Multivariate Analysis , volume =. 1973 , doi =

1973
[17]

Pacific Journal of Mathematics , volume =

Maurice Sion , title =. Pacific Journal of Mathematics , volume =. 1958 , doi =

1958
[18]

arXiv preprint arXiv:1410.3639 , year =

Fuyuhiko Tanaka , title =. arXiv preprint arXiv:1410.3639 , year =. doi:10.48550/arXiv.1410.3639 , eprint =

doi arXiv
[19]

Journal of the American Statistical Association , volume =

Wassily Hoeffding , title =. Journal of the American Statistical Association , volume =. 1963 , doi =

1963
[20]

2018 , doi =

John Watrous , title =. 2018 , doi =

2018
[21]

Electronic Journal of Probability , volume =

Sjoerd Dirksen , title =. Electronic Journal of Probability , volume =. 2015 , doi =. 1309.3522 , archiveprefix =

2015 arXiv
[22]

arXiv preprint arXiv:2510.07788 , year =

Angelos Pelecanos and Jack Spilecki and John Wright , title =. arXiv preprint arXiv:2510.07788 , year =. doi:10.48550/arXiv.2510.07788 , eprint =

doi
[23]

arXiv preprint arXiv:2602.21510 , year =

Hyukgun Kwon and Seok Hyung Lie and Liang Jiang , title =. arXiv preprint arXiv:2602.21510 , year =. doi:10.48550/arXiv.2602.21510 , eprint =

doi arXiv
[26]

Theory of Computing , volume =

Sanjeev Arora and Elad Hazan and Satyen Kale , title =. Theory of Computing , volume =. 2012 , doi =

2012
[27]

Journal of Combinatorial Theory, Series A , volume =

Norbert Sauer , title =. Journal of Combinatorial Theory, Series A , volume =. 1972 , doi =

1972
[28]

40th Annual Symposium on Foundations of Computer Science , pages =

Ashwin Nayak , title =. 40th Annual Symposium on Foundations of Computer Science , pages =. 1999 , doi =. quant-ph/9904093 , archiveprefix =

1999 arXiv
[29]

Tsybakov , title =

Alexandre B. Tsybakov , title =. 2009 , doi =

2009
[30]

Zeitschrift f

Jean Bretagnolle and Catherine Huber , title =. Zeitschrift f. 1979 , doi =

1979
[31]

Wald, Abraham , title =
[32]

, title =

Blackwell, David and Girshick, Meyer A. , title =
[33]

L\'evy, Paul , title =
[34]

Kolmogorov, A. N. , title =. Annales de l'Institut Henri Poincar\'e , volume =
[35]

Hengartner, Walter and Theodorescu, Radu , title =
[36]

Bulletin de la Soci\'et\'e Math\'ematique de France , volume =

Moreau, Jean-Jacques , title =. Bulletin de la Soci\'et\'e Math\'ematique de France , volume =. 1965 , doi =

1965
[37]

and Combettes, Patrick L

Bauschke, Heinz H. and Combettes, Patrick L. , title =. 2017 , doi =

2017
[38]

Drucker, Harris and Burges, Christopher J. C. and Kaufman, Linda and Smola, Alex and Vapnik, Vladimir , title =. Advances in Neural Information Processing Systems 9 , pages =
[39]

2026 , eprint=

Online Shadow Tomography Matching the Classical Bounds , author=. 2026 , eprint=

2026
[40]

2019 , eprint=

Gentle Measurement of Quantum States and Differential Privacy , author=. 2019 , eprint=

2019
[41]

Nature Physics , volume =

Huang, Hsin-Yuan and Kueng, Richard and Preskill, John , title =. Nature Physics , volume =. 2020 , doi =. 2002.08953 , archivePrefix =

2020 arXiv
[42]

2024 IEEE 65th Annual Symposium on Foundations of Computer Science (FOCS) , pages =

Chen, Sitan and Gong, Weiyuan and Ye, Qi , title =. 2024 IEEE 65th Annual Symposium on Foundations of Computer Science (FOCS) , pages =. 2024 , doi =. 2404.19105 , archivePrefix =

2024 arXiv
[43]

PRX Quantum , volume =

King, Robbie and Gosset, David and Kothari, Robin and Babbush, Ryan , title =. PRX Quantum , volume =. 2025 , doi =. 2404.19211 , archivePrefix =

2025 arXiv
[44]

PRX Quantum , volume =

Huang, Jerry and Lewis, Laura and Huang, Hsin-Yuan and Preskill, John , title =. PRX Quantum , volume =. 2026 , doi =. 2410.15501 , archivePrefix =

2026 arXiv
[45]

Quantum , volume =

Grier, Daniel and Pashayan, Hakop and Schaeffer, Luke , title =. Quantum , volume =. 2024 , doi =. 2211.11810 , archivePrefix =

2024 arXiv
[46]

Efficient Shadow Tomography of Thermal States , year =

Chen, Chi-Fang and Gily. Efficient Shadow Tomography of Thermal States , year =. 2603.16845 , archivePrefix =
[47]

2026 , eprint=

Instance-optimal high-precision shadow tomography with few-copy measurements: A metrological approach , author=. 2026 , eprint=

2026

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

Proceedings of the 50th Annual ACM SIGACT Symposium on Theory of Computing , series =

Scott Aaronson , title =. Proceedings of the 50th Annual ACM SIGACT Symposium on Theory of Computing , series =. 2018 , doi =. 1711.01053 , archiveprefix =

arXiv 2018

[2] [2]

Improved quantum data analysis

Costin B. Improved Quantum Data Analysis , journal =. 2024 , doi =. 2011.10908 , archiveprefix =

work page Pith review arXiv 2024

[3] [3]

Quantum Event Learning and Gentle Random Measurements

Adam Bene Watts and John Bostanci , title =. 15th Innovations in Theoretical Computer Science Conference (ITCS 2024) , series =. 2024 , doi =. 2210.09155 , archiveprefix =

work page Pith review arXiv 2024

[4] [4]

Dimension Independent and Computationally Efficient Shadow Tomography

Pulkit Sinha , title =. Proceedings of the 57th Annual ACM Symposium on Theory of Computing , series =. 2025 , doi =. 2411.01420 , archiveprefix =

work page Pith review arXiv 2025

[5] [5]

arXiv preprint arXiv:2407.13874 , year =

Sitan Chen and Jerry Li and Allen Liu , title =. arXiv preprint arXiv:2407.13874 , year =. doi:10.48550/arXiv.2407.13874 , eprint =

arXiv doi:10.48550/arxiv.2407.13874

[6] [6]

Proceedings of Thirty Ninth Conference on Learning Theory , series =

Senrui Chen and Weiyuan Gong and Sisi Zhou , title =. Proceedings of Thirty Ninth Conference on Learning Theory , series =. 2026 , url =. 2602.04952 , archiveprefix =

arXiv 2026

[7] [7]

Gentle Measurement of Quantum States and Differential Privacy

Scott Aaronson and Guy N. Rothblum , title =. Proceedings of the 51st Annual ACM SIGACT Symposium on Theory of Computing , series =. 2019 , doi =. 1904.08747 , archiveprefix =

work page Pith review arXiv 2019

[8] [8]

Sequential measurements, disturbance and property testing

Aram W. Harrow and Cedric Yen-Yu Lin and Ashley Montanaro , title =. Proceedings of the Twenty-Eighth Annual ACM-SIAM Symposium on Discrete Algorithms , pages =. 2017 , doi =. 1607.03236 , archiveprefix =

work page Pith review arXiv 2017