Latest from ICANN: Quantum Computers + N85 Peering Forum

Fri Mar 18 13:30:22 UTC 2022

Nanog News wrote:

> Latest from ICANN: Quantum Computers are "Interesting"…
> But Don't Lose your Head

But, quantum computers are mocked up by theoretical
physicists, IT amateurs who don't know basics of
computational and/or information theory at all, and,
as such, just do not work better than classical ones.

As for adiabatic quantum computers (those of D-wave),
they can be configured to form various quantum circuits.
Thus, a quantum circuit, the minimum energy state of which
is a solution to some optimization problem, may be
constructed. In addition, according to quantum mechanical
adiabatic theorem, the state can be reached within
polynomial time (w.r.t. problem size), *IF* (very big
if) energy difference between the lowest and the
second lowest energy states is not very small.

However, it is obvious that errors to construct the
circuit will badly affect the result. IIRC, D-wave
quantum computer has 8bit ADC to control coupling
strength between qubits, which is poorly accurate.

Worse, there are known classical, not quantum,
approximation algorithms to compute the lowest
energy state with certain accuracy (w.r.t. energy)
within polynomial time.

As such, "if energy difference between the
lowest and the second lowest energy states
is not very small", the classical algorithms
can find the correct answer and quantum
computing is no better than the classical
algorithms.

It should be noted that, with really hard problems,
the energy differences are very small (become
exponentially smaller as problem size increases),
which can not be solved by classical approximation
nor quantum algorithms in polynomial time.

Though it is possible to construct quantum circuits
to enhance energy differences a lot, such circuits,
like high Q resonators, require extreme accuracy,
which is impossible to construct for large problem
size, which means such attempts do not contribute
scalability.

Uselessness of quantum logic gate style quantum computers
will be discussed in a separate mail.

				Masataka Ohta