THE QUANTUM INTELLIGENCER

A paper posted to the arXiv reports an optimal algorithm for the quantum linear systems problem, according to its abstract. The paper's claim is that the query complexity of the problem is Θ(κ√d log(1/ε)), a figure that matches the known lower bounds and therefore closes the asymptotic question; the same work says it resolves an open problem of Berry and Childs, by showing that any N×N unitary can be implemented with bounded error using O(√N) queries to its matrix entries. The matter is one for the record. No committee has acted on the result, no standard is altered, and no authority is bound by it. The paper carries the authority of its argument alone; a theorem about query counts is not yet a machine in service. The abstract supplies the numbers by which the result is to be measured. In its notation, κ denotes the condition number of the matrix, d its sparsity, and ε the permitted approximation error. It records the previous best known query complexities for the linear systems problem as O(κd log(1/ε)) and κ√d(κd/ε)^{o(1)}, against lower bounds of Ω(κ log(1/ε)) and Ω(κ√d). The new figure, Θ(κ√d log(1/ε)), is the point at which the upper and lower bounds meet. On the separate unitary question, the stated figure is O(√N) queries to the matrix entries of an N×N unitary. These quantities are the record as the abstract gives it, and no further source is cited for them. The original context of these figures is the query model, a counting convention in which the cost of an algorithm is measured by the number of times it may consult an oracle rather than by revolutions of a clock. The abstract states its results in that model, and the claims carry no further institutional freight. No implementation is described, no vendor is named, no standards body is cited, and no deadline follows from the posting. What the optimal bound does is set the terms in which any subsequent claim in this line will be judged: an algorithm that exceeded the new figure, or could not match it, would be settling for something less than the abstract now says is possible. —Elias Hartwell Dispatch from The Prepared E0