New Lower-bounds for Quantum Computation with Non-Collapsing Measurements
David Miloschewsky, Supartha Podder
TL;DR
This paper analyzes the power of non-collapsing measurements in quantum computation by studying the PDQP model. It introduces an equivalent circuit-based definition to enable rigorous lower-bounding using a positive weighted adversary method and derives a quantitative trade-off between the number of queries $Q$ and non-collapsing measurements $P$, yielding bounds such as $QP = \,\Omega\(\sqrt{mm'/ll'}\)$ and, in the non-adaptive setting, $QP = \,\Omega\(\max\{m/l, m'/l'\}\)$. Applying these bounds to unstructured search, majority, parity, collision, and element distinctness, the authors obtain tight or near-tight results, including a tight $\Theta\(N^{1/3}\)\) bound for search in PDQP and $\Theta\(\sqrt{N}\)$ in the non-adaptive variant, while showing constant-time results for collision and size-bounded behavior for element distinctness. They further explore PDQP under query restrictions and compare with copy-enabled models (CBQP), highlighting the distinct advantages provided by non-collapsing measurements versus copying. The work advances a framework for understanding quantum speedups under non-standard measurement capabilities and identifies several open questions, including the adaptation of lower-bound techniques and polynomial-method approaches to PDQP. Overall, the paper clarifies the limited speedups PDQP offers over BQP and maps the landscape of lower bounds across multiple problem families in this metaphysical quantum setting.
Abstract
Aaronson, Bouland, Fitzsimons and Lee introduced the complexity class PDQP (which was original labeled naCQP), an alteration of BQP enhanced with the ability to obtain non-collapsing measurements, samples of quantum states without collapsing them. Although PDQP contains SZK, it still requires $Ω(N^{1/4})$ queries to solve unstructured search. We formulate an alternative equivalent definition of PDQP, which we use to prove the positive weighted adversary lower-bounding method, establishing multiple tighter bounds and a trade-off between queries and non-collapsing measurements. We utilize the technique in order to analyze the query complexity of the well-studied majority and element distinctness problems. Additionally, we prove a tight $Θ(N^{1/3})$ bound on search. Furthermore, we use the lower-bound to explore PDQP under query restrictions, finding that when combined with non-adaptive queries, we limit the speed-up in several cases.
