I was looking at this paper by Faulkner-Lewkowycz-Maldacena. They give a very interesting proposal of calculating one loop (i.e, 1/N) correction to EE from computing the EE between the bulk regions. The proposal is really fascinating. But I didn't quite realize few things.
What was the motivation behind this proposal? I mean why should one expect that 'this quantity' must be 1/N corrections?
What exactly one has to compute in the bulk geometry? They have some examples in the paper but I couldn't understand how to calculate the EE between bulk regions. Can someone refer to some articles where some concrete calculations have been done in this direction?