In the published version of this paper, the final equation in (A.26) on page 49 should appear as g c (τ, 0) = 6a 2,2 .(A.26)The incorrect version of this equation was implemented in the numerical solution of the system, and propagated to several places throughout the paper. The interested reader may find a revised version of the manuscript at [1]. Note that the corrected boundary condition in equation (A.26) leads to a monotonic apparent horizon area, as opposed to the nonmonotonic behaviour in the published version of figures 1 and 2 (see the appendix for the new versions). In fact, it can be shown from constraint equation (3.18) evaluated at the horizon ρ = 1 α , that the change in the apparent horizon area is given byimplying that the area of the apparent horizon monotonically increases with time, in agreement with the area theorems of [2] (see also [3]). 1
TextHere we list the required changes to the text of the paper.• Throughout the body of this paper (including the abstract) the claim is made that the thermalization times of the two-point function and entanglement entropy (EE) exceeds the thermalization time of the one-point function for wide enough separations of 2y m . However, as can be seen in the new version of figure 15 in the appendix, the new thermalization times of the two-point function and EE do not have longer thermalization times than the one-point function for the largest values of y m considered. This weakens these claims in the paper. However, as the behaviour for both observables' thermalization times is linear and monotonic, we expect that the thermalization times of the two-point function and EE will still exceed that of the one-point function for wide enough regions.• As discussed above, statements about the nonmonotonicity of the area density of the apparent horizon are no longer correct, and should be disregarded.• Because of the new table (see the next section), the discussion on page 16 should be changed from "In the table, we see that the equilibration times of the horizon become approximately constant for small α" to "In the table, we see that the equilibration times of the horizon become approximately constant for small α, (although we see some variation for the event horizon)."1 We would like to thank Mukund Rangamani and Moshe Rozali for valuable discussions regarding the area theorems.-1 -
JHEP07(2015)137• Because of the changes in figure 17 (see the appendix to this erratum), the discussion in the parentheses on page 40 should be changed from "note that in figure 17, the equilibration curve for α = 1 is not significantly later deep in the bulk than near the boundary, while the effect becomes more pronounced for the smaller values of α," to "Note that in figure 17, the equilibration curve for α = 1 is an exception, equilibrating earlier at most points in the bulk than at the boundary (even the deepest part). The effect of the integrand equilibrating later in the bulk than near the boundary is visible for the smaller values of α, which correspond more to the universal behaviou...