able to do error correction in these
systems, because it’s analog error
correction,” explained Raymond
Laflamme of the University of Waterloo and the Perimeter Institute, both
in Waterloo, Ontario.
Fortunately, those experts Laflamme
spoke of were mistaken.
AFTER DECADES OF re- search, quantum comput- ers are approaching the scale at which they could outperform their “
classical” counterparts on some problems.
They will be truly practical, however,
only when they implement quantum
error correction, which combines
many physical quantum bits, or qubits, into a logical qubit that preserves
its quantum information even when
its constituents are disrupted. Although this task once seemed impossible, theorists have developed multiple techniques for doing so, including
“surface codes” that could be implemented in an integrated-circuit-like
For ordinary binary data, errors can
be corrected, for example, using the
majority rule: A desired bit, whether
1 or 0, is first triplicated as 111 or 000.
Later, even if one of the three bits has
been corrupted, the other two “
outvote” it and allow recovery of the original data. Unfortunately, the “
no-clon-ing” theorem of quantum mechanics
forbids the duplication (or triplication) of a qubit.
Moreover, the power of quantum
computing emerges from having arbi-
trary mixtures of bit values. Since any
combination is valid, it would seem
impossible to detect a change from
the original combination, let alone
“Most people who were doing
quantum information in the ‘80s
and ‘90s would say we’ll never be
Closing in on Quantum
Quantum computers will only become practical
when they implement quantum error correction.
Science | DOI: 10.1145/3355371
“Complicated calculations fail as the systems get out of hand due to perturbations,” says
Rainer Blatt of the Institute of Experimental Physics at the University of Innsbruck and
the Institute of Quantum Optics and Quantum Information. “Using error correction, this
process can be contained.”