ABSTRACT
In its short history, the field of quantum computing has produced a num ber of truly startling revelations about the computational power inherent in quantum systems. These discoveries generated a great deal of excitement, though this excitement was tempered by the realization that the quantum correlations responsible for this power are extremely fragile. It became clear that to properly assess the performance of a quantum computer, the effects of noise and imperfect quantum gates would have to be included in the analysis. It was thought by many that the virtues of a quantum computer would not survive the presence of these effects. Remarkably, in three years (1995-1997), theoretical tools were developed that stood this expectation on its head. Be ginning with the groundbreaking work of P. W. Shor and A. M. Steane, a verygeneral framework for constructing quantum error correcting codes was estab lished, and protocols were developed that allowed quantum computations to be done in a fault-tolerant manner. This line of development culminated in a demonstration of the accuracy threshold theorem. This theorem showed that fault-tolerant protocols for quantum computation, in conjunction with concatenated quantum error correcting codes, would allow a quantum compu tation of arbitrary duration to be done with arbitrarily small error probability, provided all quantum gates used in the computation had error probabilities that fell below a value known as the accuracy threshold Pa. This threshold was calculated for a number of simple error models yielding values in the range 10-6 < p a < io -3 . The good news is that Pa ^ 0 so that reliable quantum computing in the presence of noise and imperfect quantum gates is possible. The bad news is that Pa is small so that building quantum gates that satisfy the tolerances of the threshold theorem will be technically challenging.
