Hi all,
Tomorrow's talk will be by Viktor at 5pm in the Pauli Room.
See you there,
Roger
Title: Simulating Quantum Correlations with finite Communication (RT07)
It is well known that measurements on quantum states may reveal some
behaviour which cannot be produced classically. The last 15 years people
have tried to quantify this 'non-local part' using several information
theoretic primitives such as NL-boxes or classical bits used for
communication.
In 2003 Bacon and Toner could show that 1 classical bit of one-way
communication suffices to simulate the EPR-pair setup for arbitrary
2-outcome measurements. In 2007, Regev and Toner could show that 2
classical bits of communication suffice to simulate quantum correlations
on a (known) bi-partite state of dimension d x d with 2-outcome
measurements. Surprisingly, only 2 bits of communication is needed
(independent of the dimension d). Nevertheless, even though quantum
correlations $\alpha \beta$ can be simulated, their protocol does not
yield correct marginals. Not much is known beyond these cases.
In this talk we will give a brief sketch of the techniques used by Regev
and Toner.