Stochastic inflation from a non-equilibrium renormalization group

Stochastic Inflation and Non-Equilibrium Renormalization Group

Background and Motivation

"Stochastic inflation from a non-equilibrium renormalization group" (2605.11096) addresses a pivotal problem in quantum field theory (QFT in curved spacetime and early-universe cosmology: the breakdown of perturbative methods at late times due to infrared (IR) enhancement in scalar field dynamics during inflation. Conventional stochastic inflation frameworks—often formulated as Fokker–Planck equations derived from Langevin-type models—successfully resum secular IR effects, yet their foundational connection to open-system quantum dynamics and Wilsonian RG is incomplete. The paper aims to rigorously derive the stochastic inflation description as an effective theory for super-Hubble modes via principled coarse-graining within the Schwinger-Keldysh formalism, and to elucidate how dissipative and diffusive operators emerge and are organized under non-equilibrium RG flows.

Open Effective Field Theory Construction

The authors begin with a microscopic scalar theory (e.g., $\lambda \phi^4$) and explicitly decompose the field into long and short wavelength modes across a time-dependent cutoff $\Lambda(t)$. This coarse-graining induces non-unitary evolution for the reduced density matrix of the long sector, naturally described as an open effective field theory (EFT) on the Schwinger-Keldysh contour.

Key steps:

• Influence functional derivation: Integration of short modes yields an influence functional, with coefficients determined by short-mode correlators at the coarse-graining scale. Importantly, time-locality is imposed by a narrow shell (thin-shell approximation) in momentum, while spatial locality arises dynamically in the super-Hubble regime.
• Operator hierarchy: The EFT expansion is governed by three small parameters: the shell width $\Delta\Lambda/\Lambda$, spatial gradients $p/\Lambda$, and the super-Hubble parameter $\epsilon = \Lambda/(aH)$. Systematic expansion allows controlled calculation of corrections to the leading stochastic picture.
• Operator content: The effective action for the long modes includes both dissipative (mixing branches), drift (deterministic response), and diffusive (noise) operators. At leading order in the combined expansion, diffusive terms dominate, yielding a stochastic description.

Numerical results for the Gaussian noise kernel are obtained exactly at leading order. The coefficient of the leading diffusive term $(\dot{\varphi}_a^L)^2$ is $H^3/4\pi^2$, which matches the stochastic noise amplitude in Starobinsky's original formulation.

Renormalization Group Flow for the Reduced Density Matrix

Complementary to the top-down EFT construction, the authors formulate a Wilsonian RG flow directly for the reduced density matrix, utilizing a Polchinski-type equation for the evolution as the cutoff $\Lambda$ is lowered. This RG equation generates dissipative and diffusive operators dynamically, with the effective action matching the Schwinger-Keldysh description.

Main theoretical features:

• Bilocal RG kernels: The RG flow is governed by kernels bilocal in time and space. Locality in time is enforced as the shell width becomes narrow; locality in space follows from expansion in low momenta.
• Schwinger-Keldysh structure: Functional derivatives in the RG equation generate noise ($aa$ sector) and response ($ar/ra$ sectors) components.
• Infrared hierarchy: In the super-Hubble regime ($\Lambda(t)$0), the RG flow is dominated by diffusion—the noise kernel—while drift and dissipation are suppressed by $\Lambda(t)$1. The analysis explicitly establishes the dominance of stochastic over deterministic response at late times.

The exact RG equation induces a probability distribution flow (generalized Fokker–Planck equation) for the long-wavelength modes, with corrections systematically controlled by the expansion parameters. The equation recovers the standard Fokker–Planck structure when the cutoff is in the deep infrared.

Quantum Corrections and Non-Gaussian Noise

Beyond leading order, the paper provides a rigorous calculation of controlled corrections:

• Field-dependent noise: At first order in $\Lambda(t)$2, interactions modify the two-point function of short modes, inducing a background-dependent effective mass and a multiplicative correction to the diffusion kernel. This yields field-dependent Gaussian noise amplitude.
• Non-Gaussian noise: Higher powers of the response field in the SK action correspond to higher cumulants of the stochastic force, encoded via Kramers–Moyal derivatives in the probability evolution equation. The hierarchy of these corrections is established by super-Hubble scaling, e.g., $\Lambda(t)$3 terms are suppressed by $\Lambda(t)$4.
• Operator redundancy: Detailed appendices clarify the independence and suppression of higher diffusive operators in the local expansion.

No speculative or bold claims are made about speculative regimes—corrections are explicitly organized and bounded within the parameter hierarchy.

Implications for Theory and Cosmology

The systematic link between stochastic inflation and an open-system RG framework clarifies the nature and range of validity of stochastic inflation in expanding backgrounds. The construction:

• Provides a principled origin for noise and drift terms, grounding phenomenological stochastic inflation models in open quantum system dynamics.
• Offers controlled, improvable EFT expansions for computing quantum corrections beyond the standard stochastic regime.
• Identifies the diffusive regime as the attractor of the RG flow, but notes that a true IR fixed point requires further demonstration.

The results constrain the particle interpretation of super-Hubble modes; the reduced density matrix is intrinsically mixed and the stochastic description represents the natural endpoint of the RG flow. The lack of a closed, autonomous pure-state theory at long wavelengths is quantitatively established.

Prospects for Future Development

Future developments include:

• Extension to multi-field cases, gauge interactions, and non-standard inflationary scenarios.
• Exploration of whether the diffusive regime is a genuine RG fixed point or merely an approximate attractor.
• Application of this framework to late-time cosmological phenomena and the quantum-to-classical transition.
• Quantitative analysis of corrections to cosmological observables, leveraging the systematic hierarchy detailed here.

This framework provides a solid foundation for incorporating quantum non-equilibrium effects in cosmological simulations and improves the theoretical precision of inflationary predictions, especially as higher-order and non-Gaussian corrections become observationally relevant.

Conclusion

The paper systematically reconstructs stochastic inflation as the infrared limit of an open-system Wilsonian RG flow for the reduced density matrix. The operator structure and hierarchy are derived from first principles via Schwinger-Keldysh coarse-graining and RG analysis rather than phenomenological postulates. The stochastic regime emerges as the leading diffusive infrared theory, with corrections—dissipation, drift, non-Gaussian noise—explicitly computable and parametrically suppressed. This work provides a rigorous and improvable framework for stochastic inflation, clarifies its origin and validity as an EFT, and sets the stage for controlled quantum corrections in cosmological predictions.