Comment on 'Quantum waveguide array generator for performing Fourier transforms: Alternate route to quantum computing' Appl. Phys. Lett. 79, 2823 (2001)

摘要

In their letter Akis and Ferry propose a quantum waveguide array approach for performing quantum Fourier transforms (QFTs). The waveguide produces 2~(n) waves at its output with controllable relative phases; n is the number of binary splits of the input wave. The interference pattern from these waves is recorded and implements a Fourier transform. The authors claim that their waveguide approach is "a more practical means" and an alternative to the "qubit paradigm that currently dominates the field of quantum computing" (double quotation marks are direct quotes from Ref. 1). The main result claimed by the authors is an implementation of the QFT that is as efficient as that obtained using the standard paradigm. In their conclusions they say, "... it is unclear whether the promised speedup in certain computations arises from the quantum nature of the systems or from the highly parallel analog processing that is provided by the array of qubits. We have argued that it is the latter that is important, and that equal speedup is available using analog processing arrays whose operation is based on general wave principles." The arguments leading to this conclusion are unfortunately based on an incorrect assumption: that interference is sufficient to obtain a quantum speedup. The essence of the waveguide approach is quantum interference. Indeed, the authors claim: "Given that quantum mechanics is primarily a wave mechanics concept, these examples based on electromagnetic and acoustic waves suggest that there should be a more natural approach to quantum signal processing than that found in the existing quantum computing literature."