Nonadaptive quantum query complexity

We study the power of nonadaptive quantum query algorithms, which are algorithms whose queries to the input do not depend on the result of previous queries. First, we show that any bounded-error nonadaptive quantum query algorithm that computes some total boolean function depending on n variables must make Omega(n) queries to the input in total. Second, we show that, if there exists a quantum algorithm that uses k nonadaptive oracle queries to learn which one of a set of m boolean functions it has been given, there exists a nonadaptive classical algorithm using O(k log m) queries to solve the same problem. Thus, in the nonadaptive setting, quantum algorithms can achieve at most a very limited speed-up over classical query algorithms.

💡 Research Summary

The paper investigates the computational power of non‑adaptive quantum query algorithms—algorithms that issue all their oracle queries in advance, without using the outcomes of earlier queries to decide later ones. Two complementary lower‑bound results are established, showing that in this restricted model quantum speed‑ups are at most modest.

First, the authors prove that any bounded‑error (error ≤ 1/3) non‑adaptive quantum algorithm that computes a total Boolean function f : {0,1}ⁿ → {0,1} must make at least Ω(n) oracle queries. The proof adapts the polynomial method and a spectral‑analysis of the query operator to the non‑adaptive setting. Because the entire query sequence is fixed, the overall unitary can be expressed as a single linear combination of the input bits. By examining the influence of each input variable on the final measurement, the authors show that, unless almost every variable is queried, the algorithm cannot distinguish inputs that differ on the function’s critical bits. Consequently, the linear‑size lower bound that holds for classical deterministic algorithms also holds for quantum algorithms when adaptivity is forbidden.

Second, the paper addresses a learning problem: given a set ℱ = {f₁,…,f_m} of Boolean functions, an unknown target f ∈ ℱ is hidden, and the algorithm may ask non‑adaptive quantum queries to an oracle for f. If a quantum algorithm can identify the hidden function using k non‑adaptive queries, then there exists a classical non‑adaptive algorithm that succeeds with O(k log m) queries. The argument proceeds by quantifying the information content of a single quantum query. A quantum query can extract at most O(log m) bits about which function is present, because the post‑measurement outcome lives in a space of dimension 2^k and must discriminate among m possibilities. Repeating the process k times yields O(k log m) bits of information, which a classical algorithm can use to perform a binary‑search‑type reduction of the candidate set. Thus the quantum advantage collapses to a logarithmic factor.

Together, these results demonstrate that the dramatic exponential improvements known for adaptive quantum algorithms (e.g., Grover’s search, Simon’s problem, Shor’s factoring) do not survive when adaptivity is removed. The non‑adaptive constraint prevents the algorithm from exploiting intermediate measurement results to steer later superpositions, limiting the benefit of quantum parallelism. While constant‑factor improvements may still be possible for specially structured functions, the asymptotic gap between quantum and classical query complexities in the non‑adaptive regime is at most polylogarithmic.

The authors discuss implications for hardware architectures where feedback is costly or impossible, such as highly parallel quantum processors or systems with restricted classical control loops. In such settings, designers should temper expectations of quantum speed‑ups, as the theoretical upper bound on advantage is modest. The paper thus provides a clear, rigorous baseline for evaluating non‑adaptive quantum algorithms and helps focus future research on scenarios where adaptivity can be retained or where problem structure can be leveraged to obtain genuine quantum gains.