A Differential Representation of Cosmological Wavefunctions

TL;DR

This paper introduces a new differential representation for cosmological wavefunctions in de Sitter space, simplifying calculations of tree-level wavefunction coefficients for a broad class of theories relevant to cosmology.

Contribution

It presents a novel algebraic and differential method to compute cosmological wavefunctions, reducing complex nested integrals to simpler algebraic operations for many theories.

Findings

Simplifies computation of cosmological wavefunctions

Applicable to theories like effective field theory of inflation

Reduces nested integrals to algebraic operations

Abstract

Our understanding of quantum field theory rests largely on explicit and controlled calculations in perturbation theory. Because of this, much recent effort has been devoted to improve our grasp of perturbative techniques on cosmological spacetimes. While scattering amplitudes in flat space at tree level are obtained from simple algebraic operations, things are harder for cosmological observables. Indeed, computing cosmological correlation functions or the associated wavefunction coefficients requires evaluating a growing number of nested time integrals already at tree level, which is computationally challenging. Here, we present a new "differential" representation of the cosmological wavefunction in de Sitter spacetime that obviates this problem for a large class of phenomenologically relevant theories. Given any tree-level Feynman-Witten diagram, we give simple algebraic rules to write down a seed function and a differential operator that transforms it into the desired wavefunction coefficient for any scale-invariant, parity-invariant theory of massless scalars and gravitons with general boost-breaking interactions. In particular, this applies to large classes of phenomenologically relevant theories such as those described by the effective field theory of inflation or solid inflation. Trading nested bulk time integrals for derivatives on boundary kinematical data provides a great computational advantage, especially for processes involving many vertices.