Source-linked AI summary
Quantum Computation as Geometry
Michael A. Nielsen, Mark R. Dowling, Mile Gu, Andrew C. Doherty
TL;DR
The paper asks how to characterize the difficulty of synthesizing quantum operations and constructs a geometric distance related to gate complexity. It shows that this distance supports polynomial-size approximate circuits and provides a lower bound on exact synthesis, while identifying unresolved questions about geodesics.
Problem
The paper addresses how to measure the difficulty of implementing a unitary operation and whether its geometric distance reflects the gates needed to synthesize it.
Source: p. 4Method
The paper constructs a metric from an optimal-control Hamiltonian, penalizing three- and more-body terms, and uses its geodesics to construct approximate circuits.
Source: pp. 3, 5Results
O(n^6d(I, U)^3) one- and two-qubit gates suffice to synthesize a unitary within any constant error, while d(I, U) is also a lower bound up to a constant factor on exact gate count.
Source: pp. 4, 9Takeaways & Limitations
The geometric formulation offers an alternative route to suggesting efficient quantum algorithms or proving that an algorithm is optimal.
Source: p. 9The general behavior and classification of the metric's geodesics remain unresolved, and determining the initial velocity needed to find a minimal geodesic is not generally easy.
Source: pp. 3, 5, 9
Abstract
from arXiv · showhide
Quantum computers hold great promise, but it remains a challenge to find efficient quantum circuits that solve interesting computational problems. We show that finding optimal quantum circuits is essentially equivalent to finding the shortest path between two points in a certain curved geometry. By recasting the problem of finding quantum circuits as a geometric problem, we open up the possibility of using the mathematical techniques of Riemannian geometry to suggest new quantum algorithms, or to prove limitations on the power of quantum computers.
3. For any H
The proof approximates evolution under a Hamiltonian by subdividing the interval and replacing each step with one- and two-qubit gates. It bounds the approximation error and derives a gate-count bound.
3. For any H
- The Dyson series converges in the finite-dimensional setting, and comparing it with the exponential of the mean Hamiltonian cancels the O(∆^0) and O(∆^1) terms.The proof then applies the triangle inequality and the norm inequality ∥XY∥≤∥X∥∥Y∥.Source: §3 · p. 10
- The interval [0, ∆] is divided into N = 1/∆ steps of size ∆^2 before the evolution is approximated.The proof invokes standard quantum simulation techniques.Source: §3 · p. 11
- The Hamiltonian H has L = O(n^2) one- and two-body terms, and each factor in the product is an allowed one- or two-qubit gate.The proof uses standard norm inequalities and the triangle inequality to bound the approximation error.Source: §3 · p. 11
- At most c1n^2/∆ quantum gates approximate e^−iH∆, with the error bound established using Eq. (25).The passage identifies c1 as a constant.Source: §3 · p. 11