Failure of the stochastic approach to inflation in constant-roll and ultra-slow-roll [CL]

http://arxiv.org/abs/1807.09057


After giving a pedagogical review we clarify that the stochastic approach to inflation is generically valid only at zeroth order in the (geometrical) slow-roll parameter $\epsilon_1$ if and only if $\epsilon_1\ll 6/\epsilon_2^2$, with the notable exception of slow-roll. This is due to the failure of the separate universe approach. As a byproduct, by keeping the formalism in its regime of validity, we show that quantum diffusion in ultra-slow-roll might only contribute to the power spectrum at sub-leading orders in slow-roll, and thus, it is irrelevant.

Read this paper on arXiv…

D. Cruces, C. Germani and T. Prokopec
Wed, 25 Jul 18
30/60

Comments: 16 pages