IQI Weekly Seminar

Random numbers have countless applications, and so the kind of
randomness postulated in quantum physics--"true randomness"--may be a
vital resource.  A line of research has developed whose goal is to
harness this resource.  The central goal is simple: construct a protocol
which uses quantum devices to generate bits, and at the same time,
certify that the bits are truly random.  Crucially, the certification
procedure must not depend on any prior trust in the accuracy of the devices.

Colbeck's thesis (2006) proposed a scheme for quantum random number
generation. While the scheme is simple, its security proof was very
challenging--the first (and only) previous full proof was given by
Vazirani and Vidick in 2012.

In our work we have taken several steps forward from Vazirani-Vidick
2012 and brought quantum RNG close to the point of practical
implementation.  We have constructed a security proof that is robust
(error-tolerant), that provides cryptographic security, and that is
implementable with constant quantum memory.  The proof invents multiple
new techniques which we are hopeful will find applications elsewhere.

Joint work with Yaoyun Shi.

Reference: http://arxiv.org/abs/1402.0489