Pith. Sign in

REVIEW

Quantum Query Complexity Beyond the Worst Case

Quantum Physics · Submitted 28 Sep 2026 · 127 references

Pith's verdict: A real research program is born here, but the factor of trust it earns rests on one hand-verified averaging lemma and one unverified reversible reduction; send it to a referee with both flagged as the explicit verification targets.

T0 · AI review

How this review was made: An AI referee (deepseek/deepseek-v4-flash) read the paper on September 29, 2026 and answered a fixed list of questions about it. T0 is the starting level every paper gets. The levels describe how deeply Pith has checked the paper, never the authors.

  1. T0 AI review An AI read the paper and answered a fixed list of questions.
  2. T1 Audit Automatic checks that the references and files the paper relies on exist and match.
  3. T2 Compute The paper's code or calculations re-run with the same results.
  4. T3 Formal statement The claims written out in a formal math language a computer can read.
  5. T4 Proof Those claims proved, and the proof checked by computer.

How Pith checks papers →

· 3 major objections · 4 minor · Ask about this paper · The compiled record · Read the full review →
Challenge this review Comment (0) Read on arXiv ↗
Cite ▾
arXiv ID2609.35580 (version 1)
Pith IDHGF6ZU3D, Pith's permanent record of this paper and its review
CitationAmin Shiraz Gilani, Srinivasan Arunachalam, Yanlin Chen. "Quantum Query Complexity Beyond the Worst Case." arXiv:2609.35580 (2026). https://pith.science/paper/HGF6ZU3D
Cite this reviewBibTeX for Pith's review of this paper
Claim paper

Why Pith says this

Highlight any text to comment on it.

Referee report
The paper

The paper claims

Measuring query complexity after a small random perturbation of the input reveals quantum speedups that worst-case analysis hides: the total Simon problem becomes quantum-easy while staying classically hard, and one pair of parameters…

The reading

This paper claims that measuring query complexity under small random perturbations — the smoothed-analysis model long used in classical algorithms — exposes quantum advantages that worst-case analysis hides. First, the total (promise-free) Simon problem, which needs $\Omega(2^{n/2})$ quantum queries and $\Omega(2^n)$ randomized queries in the worst case, becomes solvable in $O(n/\sigma)$ smoothed quantum queries while its smoothed classical complexity stays $\Omega(2^{n/2})$; for every $\sigma \ge 1/\mathrm{poly}(n)$ this is an exponential quantum-classical separation. Second, for every symmetric Boolean function the paper pins down both smoothed quantum and randomized complexity up to polylog factors in terms of two parameters, the transition cost and the weighted influence. Third, it gives superpolynomial smoothed quantum speedups for pattern matching and a quantum algorithm for $O(1/\sigma)$-approximate smoothed edit distance, powered by an improved quantum collision-estimation routine. The unifying message is that smoothing — perturbing each input bit with probability $\sigma$, then taking the worst case over original inputs of expected query cost — interpolates between the worst-case results of Beals et al. and the average-case results of Ambainis and de Wolf.

What carries the argument

Three mechanisms carry the argument. For Total Simon, the machine is ordinary Fourier sampling on the perturbed truth table: with probability at least $1 - 2^{-\Omega(n)}$ the $\sigma$-perturbation leaves at least $N\sigma/64$ singleton fibers, each measurement of a singleton fiber yields a Fourier sample uniform over $\mathbb{F}_2^n$, and $t = \Theta(n/\sigma)$ samples certify the absence of a period unless a rare fallback queries the full table. For symmetric functions, the load-bearing objects are the two parameters $\mathrm{transcost}_\sigma(f)$ and $w\mathrm{Inf}_\sigma(f)$, built from the transition set $T(f)$ of adjacent Hamming weights where the value changes, together with Lemma 4.5, an averaging lemma that converts per-weight costs into these two forms: it uses log-concavity of the perturbed-weight distribution and a partition of the Hamming layers into maximal constancy intervals, so that nearby transitions contribute through their total interval probability and endpoint contributions accumulate into the weighted influence. For edit distance, the machinery is a quantum walk on Johnson graphs yielding a gap-sensitive collision tester with $\tilde{O}\big(\big(\frac{n\sqrt{\ell_2}}{\ell_2 - \ell_1}\big)^{2/3}\big)$ queries, bootstrapped to multiplicative and additive collision estimators (improving the $\varepsilon$ dependence from $\varepsilon^{-25/24}$ to $\varepsilon^{-2/3}$), a variable-time approximate-counting primitive, and a reversible subset database that gives oracle access to the implicit Ulam instance while preserving the locality and separation properties of the Andoni–Krauthgamer reduction. The authors disclose that Lemma 4.5 was produced with GPT assistance and hand-verified; there is no machine-checked proof of it or of the reversible reduction.

Extended reading notes

Core claim

The central claim, on the authors' own terms, is twofold. Theorem 3.2 with Lemma 3.4 establishes that there exists a total function — the total version of Simon's problem with no promise — whose smoothed quantum query complexity is $O(n/\sigma)$, exponentially smaller than its smoothed classical query complexity $\Omega(2^{n/2})$, for every $\sigma \ge 1/\mathrm{poly}(n)$; the algorithm is the standard Fourier-sampling primitive for hidden subgroups, and smoothing creates enough singleton fibers to make each Fourier sample uniform and the absence of the period certifiable. Theorem 4.1 then characterizes every total symmetric Boolean function: the smoothed randomized complexity is $\tilde{\Theta}(\min\{n, \mathrm{transcost}_\sigma(f)^2\})$ and the smoothed quantum complexity is $\tilde{\Theta}(w\mathrm{Inf}_\sigma(f) + \mathrm{transcost}_\sigma(f))$, where the transition cost measures the hardest single weight-distinction after accounting for the spread of the perturbed weight and the weighted influence sums per-transition costs weighted by the probabilities of the adjacent Hamming layers. At $\sigma = 0$ this recovers the Beals et al. worst-case characterization and at $\sigma = 1$ the Ambainis–de Wolf average-case results, so the two parameters unify previously separate regimes. For strings, the paper proves superpolynomial smoothed quantum speedups for pattern matching and, combining a new gap-sensitive and multiplicative collision estimator with a reversible implementation of the Andoni–Krauthgamer reduction, a quantum algorithm for $O(1/\sigma)$-approximate smoothed edit distance in $\tilde{O}\big(\frac{1}{\sigma}\big(\frac{n}{R} + \frac{n^{2/3}}{R^{1/3}}\big)\big)$ queries, where $R$ is the edit distance; for $R = \Theta(n)$ this is $\tilde{O}(n^{1/3}/\sigma)$, beating the classical $\tilde{O}(\sqrt{n}/\sigma)$.

Load-bearing premise

The deepest result, Theorem 4.1, collapses if Lemma 4.5 fails: the lemma, disclosed as GPT-assisted and hand-verified rather than machine-checked, is what converts per-Hamming-weight query costs into the $\tilde{\Theta}(\mathrm{transcost}, w\mathrm{Inf})$ forms, and a subtle error in its anti-concentration or interval-partition argument would shift the characterization. The edit-distance result additionally stands on the reversible implementation of the Andoni–Krauthgamer reduction preserving the locality and separation properties needed for the implicit Ulam instance.

Editorial extensions

If this is right

  • For every $\sigma \ge 1/\mathrm{poly}(n)$ there is a total Boolean function whose smoothed quantum complexity is exponentially smaller than its smoothed classical complexity, so the worst-case polynomial relation between quantum and randomized query complexity for total functions does not survive smoothing.
  • Theorem 4.1 unifies the field's two classic regimes: $\sigma = 0$ recovers the Beals et al. worst-case characterization and $\sigma = 1$ recovers the Ambainis–de Wolf average-case results for symmetric functions, with intermediate $\sigma$ interpolating between them.
  • Under full smoothing, Majority becomes $\tilde{\Theta}(\sqrt{n})$ for quantum queries while remaining $\tilde{\Theta}(n)$ for randomized queries — a polynomial quantum speedup that appears only in the smoothed model.
  • The $O(1/\sigma)$-approximate smoothed edit distance algorithm runs in $\tilde{O}\big(\frac{1}{\sigma}\big(\frac{n}{R} + \frac{n^{2/3}}{R^{1/3}}\big)\big)$ quantum queries, improving the classical $\tilde{O}\big(\frac{1}{\sigma}\big(\frac{n}{R} + \sqrt{n}\big)\big)$ for $R = \omega(\sqrt{n})$, and giving $\tilde{O}(n^{1/3}/\sigma)$ when $R = \Theta(n)$.
  • The improved collision estimator, with $\varepsilon$ dependence dropping from $\varepsilon^{-25/24}$ to $\varepsilon^{-2/3}$, is a transferable primitive for sublinear quantum algorithms beyond edit distance.

Reading between the lines

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

  • The Total Simon result suggests a general design principle that the paper does not itself state: smoothing may turn standard quantum primitives (Fourier sampling, amplitude estimation) that are exponentially bad in the worst case into efficient procedures, because the noise destroys the fragile structure that created worst-case hardness; a natural test is non-abelian Fourier sampling on smoothed g
  • Smoothed analysis could serve as a principled middle ground between worst-case and average-case complexity for judging when quantum advantage is 'realistic', and extending it to heuristic quantum algorithms such as QAOA might yield rigorous guarantees where none currently exist — another question the authors leave open.
  • If Lemma 4.5 holds, the two-parameter characterization implies that smoothed complexity of symmetric functions is computable in polynomial time for any given $\sigma$, making 'how hard is this function under noise' a tractable query — a computational tool the paper does not itself exploit.
Original abstract

Smoothed analysis is a central framework in classical algorithms for explaining the performance of algorithms beyond the worst case, often explaining why algorithms perform well in practice. We initiate a systematic study of its quantum counterpart and show the following results. $(1)$ We show that there is a total function whose smoothed quantum query complexity is exponentially smaller than its classical query complexity. $(2)$ We give near-tight characterizations of smoothed randomized and quantum query complexities for symmetric Boolean functions, unifying the worst-case complexity results of [Beals et al, FOCS'98] and average-case complexity results of [Ambainis and de Wolf, STACS'00]. $(3)$ We study string problems such as pattern matching and edit distance and, in various regimes, give polynomial to superpolynomial quantum speedups. Our main technical ingredients include a near-tight quantum algorithm for $\varepsilon$-approximating the number of collisions between two non-repetitive strings, improving the result of Le Gall and Ng [QIC'22]. Together, our results show that smoothing can reveal larger quantum speedups than worst-case analysis suggests, opening a path towards quantum advantage on more realistic inputs.

The compiled record

The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.

Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-09-29: 1 request, 5,706 input tokens.

What the abstract itself states, as probabilities: claims first 0.57 · states limitation 0.09 · human participants 0.01 · controlled comparison 0.15 · randomized assignment 0.02 · states quantitative result 0.63 · compares with prior results 0.95 · mentions code or data availability 0.02.

The abstract, sentence by sentence

arXiv listing abstract v1. Results and the main claim are underlined; .

Smoothed analysis is a central framework in classical algorithms for explaining the performance of algorithms beyond the worst case, often explaining why algorithms perform well in practice. We initiate a systematic study of its quantum counterpart and show the following results. $(1)$ We show that there is a total function whose smoothed quantum query complexity is exponentially smaller than its classical query complexity. $(2)$ We give near-tight characterizations of smoothed randomized and quantum query complexities for symmetric Boolean functions, unifying the worst-case complexity results of [Beals et al, FOCS'98] and average-case complexity results of [Ambainis and de Wolf, STACS'00]. $(3)$ We study string problems such as pattern matching and edit distance and, in various regimes, give polynomial to superpolynomial quantum speedups. Our main technical ingredients include a near-tight quantum algorithm for $\varepsilon$-approximating the number of collisions between two non-repetitive strings, improving the result of Le Gall and Ng [QIC'22]. Together, our results show that smoothing can reveal larger quantum speedups than worst-case analysis suggests, opening a path towards quantum advantage on more realistic inputs.

background contribution method result interpretation main claim

What the classifier reports for each sentence. The number is the probability it gave the label.
#sentencerolethe authors' own workstates the main conclusionstrength
s0 Smoothed analysis is a central framework in classical algorithms for explaining the performance of algorithms beyond the worst case, often explaining why algorithms perform well in practice. background 1.00 0.03 0.02 asserted 0.99
s1 We initiate a systematic study of its quantum counterpart and show the following results. contribution 0.99 0.81 0.05 asserted 0.94
s2 $(1)$ We show that there is a total function whose smoothed quantum query complexity is exponentially smaller than its classical query complexity. result 1.00 0.96 0.22 asserted 1.00
s3 $(2)$ We give near-tight characterizations of smoothed randomized and quantum query complexities for symmetric Boolean functions, unifying the worst-case complexity results of [Beals et al, FOCS'98] and average-case complexity results of [Ambainis and de Wolf, STACS'00]. result 0.96 0.92 0.09 asserted 1.00
s4 $(3)$ We study string problems such as pattern matching and edit distance and, in various regimes, give polynomial to superpolynomial quantum speedups. result 0.87 0.96 0.06 asserted 1.00
s5 Our main technical ingredients include a near-tight quantum algorithm for $\varepsilon$-approximating the number of collisions between two non-repetitive strings, improving the result of Le Gall and Ng [QIC'22]. method 0.76 0.93 0.04 asserted 1.00
s6 Together, our results show that smoothing can reveal larger quantum speedups than worst-case analysis suggests, opening a path towards quantum advantage on more realistic inputs. interpretation 1.00 0.81 0.93 asserted 0.92

What kind of paper the classifier reports this is: theoretical result 0.99 · evidence from proof or derivation 1.00.

What was examined, and what was not

sources read
  • arXiv listing abstract v1
not examined
  • full text
  • figures and tables
  • references and citing sentences
  • relations to other papers
levels run
  • P abstract
not run
  • S sections
  • T body sentences
  • C citations
  • R cross-paper relations

Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. Every number in the measurements table is the authors', checked by code. Record as JSON

References

Discussion

Disagree with the review, or know something it missed? Leave a comment, highlight any passage above to comment on it, or challenge this review formally.