Differential topology of adiabatically controlled quantum processes (Q352700)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Differential topology of adiabatically controlled quantum processes
scientific article

    Statements

    Differential topology of adiabatically controlled quantum processes (English)
    0 references
    0 references
    0 references
    0 references
    5 July 2013
    0 references
    The paper under review provides an interesting alternative viewpoint on adiabatic quantum algorithms. Whereas many studies analyse the spectral properties of the adiabatic Hamiltonian interpolating between Hamiltonian \(H_0\) and \(H_1\) directly, the authors use previous results linking these to critical points of the quadratic map \(H_0 + \text{i} H_1\). This alternative approach could be of large potential use to understand and perhaps optimize the performance of adiabatic quantum computation, for practical applications the evaluation of the quadratic map appears simpler than the diagonalization of the problem. After a brief exposition of the theory [\textit{E.A. Jonckheere}, \textit{F. Ahmad} and \textit{E. Gutkin}, Linear Algebra Appl. 279, No. 1-3, 227--254 (1998; Zbl 0934.15030); \textit{E. Gutkin}, \textit{E.A. Jonckheere} and \textit{M. Karow}, Linear Algebra Appl. 376, 143--171 (2004; Zbl 1050.15026)], the authors illustrate their method for Hamiltonians that can be written as the sum of two projectors, e.g. the adiabatic version of Grovers algorithm. They show that the conventional results are recovered with their method. In section 4 it is shown that some properties of the map for such Hamiltonians (that have many degeneracies due to symmetries) however are extremely vulnerable to perturbations. Nevertheless, the main result that cusps and swallow tails of the associated quadratic map occur near the avoided crossings persists. The authors then proceed to the Ising chain with a transverse field in both \(x\) and \(z\) directions rendering it an NP-complete problem. It would have been interesting to compare with the analytically tractable case of a vanishing field in \(z\)-direction. Finally, the authors apply their method to the adiabatic random walk. The authors conclude that upon detection of swallow tails or cusps one might modify the adiabatic path to avoid small gaps. The paper is dense but well-written and relatively self-contained. Readers should however be familiar with the basic principles of adiabatic computation or with at least one of the models considered.
    0 references
    0 references
    adiabatic theorem
    0 references
    numerical range
    0 references
    cusps
    0 references
    swallow tails
    0 references
    gap
    0 references
    anti-crossing
    0 references
    0 references