Lucien Heurtier
@lucienheurtier.bsky.social
Theoretical Physicist at King's Coll. London @KCL_Physics #cosmology, #particles, #Universe, #darkmatter, #blackholes. Write fast, the Universe expands...
https://lheurtier.github.io/index.html
https://lheurtier.github.io/index.html
I'm not surprised. Some of my research collaborators do the same. People also don't talk to people on the street, they date online, no? 😅 I am pretty sure LLMs are going to sell at some point academic programs and maybe even degrees. I don't see how this is not going to happen. I hope I'm wrong.
November 21, 2025 at 7:46 AM
I'm not surprised. Some of my research collaborators do the same. People also don't talk to people on the street, they date online, no? 😅 I am pretty sure LLMs are going to sell at some point academic programs and maybe even degrees. I don't see how this is not going to happen. I hope I'm wrong.
The paper can be downloaded here: arxiv.org/abs/2511.05296
Thanks again to my great collaborators Jean Alexandre and Silvia Pla!
Thanks again to my great collaborators Jean Alexandre and Silvia Pla!
Exact Renormalisation Group Evolution of the Inflation Dynamics: Reconciling $α$-Attractors with ACT
We present a non-perturbative framework for the dynamics of slow-roll inflation that consistently incorporates quantum corrections, based on an alternative functional renormalisation group (RG) approa...
arxiv.org
November 10, 2025 at 12:31 PM
The paper can be downloaded here: arxiv.org/abs/2511.05296
Thanks again to my great collaborators Jean Alexandre and Silvia Pla!
Thanks again to my great collaborators Jean Alexandre and Silvia Pla!
Coarse Graining in the context of inflation or de Sitter is not new. However, such a dynamical treatment of the potential flow is, and we hope this will open a new pathway towards understanding the effect of renormalisation on interpreting CMB data.
November 10, 2025 at 12:31 PM
Coarse Graining in the context of inflation or de Sitter is not new. However, such a dynamical treatment of the potential flow is, and we hope this will open a new pathway towards understanding the effect of renormalisation on interpreting CMB data.
We applied this paradigm to the case of alpha-attractor E-models, which were recently disfavoured by CMB observations, in particular results from the Atacama Cosmology Telescope. What we fine is remarkable: accounting for this effect rescues alpha-attractors!
November 10, 2025 at 12:31 PM
We applied this paradigm to the case of alpha-attractor E-models, which were recently disfavoured by CMB observations, in particular results from the Atacama Cosmology Telescope. What we fine is remarkable: accounting for this effect rescues alpha-attractors!
Starting with initial de Sitter initial conditions, we find that the effect of the running increases with the growing curvature of the potential. In return, the negative curvature also grows, destabilising the slow-roll attractor much earlier than expected!
November 10, 2025 at 12:31 PM
Starting with initial de Sitter initial conditions, we find that the effect of the running increases with the growing curvature of the potential. In return, the negative curvature also grows, destabilising the slow-roll attractor much earlier than expected!
Taken together, the equations of motion for the potential flow, the metric and the inflaton field provide us a master system of equations (primes denote derivative wrt the number of e-folds) that we are to solve numerically:
November 10, 2025 at 12:31 PM
Taken together, the equations of motion for the potential flow, the metric and the inflaton field provide us a master system of equations (primes denote derivative wrt the number of e-folds) that we are to solve numerically:
Well, but if the Lagrangian is a function of H(t) and phi(t), the Friedmann equations also need to be upgraded. It gives this:
November 10, 2025 at 12:31 PM
Well, but if the Lagrangian is a function of H(t) and phi(t), the Friedmann equations also need to be upgraded. It gives this:
We then use a trick: rather than trying to evaluate the inverse of the D_E operator, we multiply both sides of this equation by D_E, providing us with an equation of motion for the potential flow, which looks like that:
November 10, 2025 at 12:31 PM
We then use a trick: rather than trying to evaluate the inverse of the D_E operator, we multiply both sides of this equation by D_E, providing us with an equation of motion for the potential flow, which looks like that:
Using the Local Potential Approximation (LPA), and a top hat regulator form, we obtain the following flow equation for the inflation potential U(phi,H):
November 10, 2025 at 12:31 PM
Using the Local Potential Approximation (LPA), and a top hat regulator form, we obtain the following flow equation for the inflation potential U(phi,H):
In a time-independent system, this cutoff scale can be varied from infinity (the purely classical limit) to zero (the full quantum theory), and the flow equation integrated over the flow, to recover one limit from the other.
November 10, 2025 at 12:31 PM
In a time-independent system, this cutoff scale can be varied from infinity (the purely classical limit) to zero (the full quantum theory), and the flow equation integrated over the flow, to recover one limit from the other.
In a time-independent system, this cutoff scale can be varied from infinity (the purely classical limit) to zero (the full quantum theory), and the flow equation integrated over the flow, to recover one limit from the other.
November 10, 2025 at 12:31 PM
In a time-independent system, this cutoff scale can be varied from infinity (the purely classical limit) to zero (the full quantum theory), and the flow equation integrated over the flow, to recover one limit from the other.
In this set-up, the path integral leading to the calculation of the effective action is regulated to only incorporate quantum modes with frequencies larger than a cutoff scale k. Then, an exact flow equation of the effective action with respect to k is derived.
November 10, 2025 at 12:31 PM
In this set-up, the path integral leading to the calculation of the effective action is regulated to only incorporate quantum modes with frequencies larger than a cutoff scale k. Then, an exact flow equation of the effective action with respect to k is derived.
Fortunately, there exists a formalism, introduced and developed over the years by Wetterich, Litim, and others, that is particularly appropriate to discuss this: the Functional Renormalisation Group (or Exact Renormalisation Group).
November 10, 2025 at 12:31 PM
Fortunately, there exists a formalism, introduced and developed over the years by Wetterich, Litim, and others, that is particularly appropriate to discuss this: the Functional Renormalisation Group (or Exact Renormalisation Group).
That's exactly the question we ask in this paper: Can a scalar field theory remain purely classical while sub-horizon quantum modes constantly exit the horizon, hence effectively "classicalizing".
November 10, 2025 at 12:31 PM
That's exactly the question we ask in this paper: Can a scalar field theory remain purely classical while sub-horizon quantum modes constantly exit the horizon, hence effectively "classicalizing".
Yes, in general relativity, constant energy density and constant de Sitter horizon don't imply that all quantum modes contribute in the exact same way to an effective field theory over time: the wavelength of every mode redshifts with expansion and UV modes become IR.
November 10, 2025 at 12:31 PM
Yes, in general relativity, constant energy density and constant de Sitter horizon don't imply that all quantum modes contribute in the exact same way to an effective field theory over time: the wavelength of every mode redshifts with expansion and UV modes become IR.
But wait wait wait: The Hubble scale is quite constant during inflation, but many modes become super-horizon. In the comoving frame the UV/IR cutoff is clearly varying with time? Does this mean that some modes contribute to the effective inflation sometimes but not always?
November 10, 2025 at 12:31 PM
But wait wait wait: The Hubble scale is quite constant during inflation, but many modes become super-horizon. In the comoving frame the UV/IR cutoff is clearly varying with time? Does this mean that some modes contribute to the effective inflation sometimes but not always?
But wait, inflation potentials favoured by Planck/BICEP/Keck are all concave, no? Yes, that's fine, because not all quantum modes contribute to the path integral in that case. Only those that are sub-horizon. Super-horizon modes are IR/classical modes.
November 10, 2025 at 12:31 PM
But wait, inflation potentials favoured by Planck/BICEP/Keck are all concave, no? Yes, that's fine, because not all quantum modes contribute to the path integral in that case. Only those that are sub-horizon. Super-horizon modes are IR/classical modes.
In QFT, obtaining the effective potential for a classical background field requires performing a path integral over all quantum perturbations. Starting from an arbitrary potential, this typically leads to an effective potential that is convex.
November 10, 2025 at 12:31 PM
In QFT, obtaining the effective potential for a classical background field requires performing a path integral over all quantum perturbations. Starting from an arbitrary potential, this typically leads to an effective potential that is convex.
During inflation, the universe is approximately de Sitter. The inflaton typically rolls along a potential classically. Nobody really cares about asking "at what scale is this potential defined?" because, well, during inflation, energy scales don't evolve that much after all.
November 10, 2025 at 12:31 PM
During inflation, the universe is approximately de Sitter. The inflaton typically rolls along a potential classically. Nobody really cares about asking "at what scale is this potential defined?" because, well, during inflation, energy scales don't evolve that much after all.
Thanks again to my great collaborators Malcolm Fairbairn and María Olalla Olea-Romacho! Please go read the paper there: arxiv.org/pdf/2511.01612
and feel free to use our Potential Inflation Posterior Emulator (PIPE), which you can download here: gitlab.com/cosmoPipe/pi...
and feel free to use our Potential Inflation Posterior Emulator (PIPE), which you can download here: gitlab.com/cosmoPipe/pi...
arxiv.org
November 4, 2025 at 9:36 AM
Thanks again to my great collaborators Malcolm Fairbairn and María Olalla Olea-Romacho! Please go read the paper there: arxiv.org/pdf/2511.01612
and feel free to use our Potential Inflation Posterior Emulator (PIPE), which you can download here: gitlab.com/cosmoPipe/pi...
and feel free to use our Potential Inflation Posterior Emulator (PIPE), which you can download here: gitlab.com/cosmoPipe/pi...
Naturally, our findings do not overrule other non-primordial interpretations of the eBOSS tension, such as axion-like or warm dark matter components, suppressing the matter power spectrum on small scales. But it provides a compelling framework to understand it from a cosmic inflation point of view.
November 4, 2025 at 9:36 AM
Naturally, our findings do not overrule other non-primordial interpretations of the eBOSS tension, such as axion-like or warm dark matter components, suppressing the matter power spectrum on small scales. But it provides a compelling framework to understand it from a cosmic inflation point of view.
To put it in a nutshell, potentials with features manage to fit the data incredibly better than featureless potentials. We now have the pipeline in place to analyse even more datasets, and our code is publicly available here: gitlab.com/cosmoPipe/pi...
PIPE / PIPE – Potential Inflation Posterior Emulator · GitLab
GitLab.com
gitlab.com
November 4, 2025 at 9:36 AM
To put it in a nutshell, potentials with features manage to fit the data incredibly better than featureless potentials. We now have the pipeline in place to analyse even more datasets, and our code is publicly available here: gitlab.com/cosmoPipe/pi...
Such localised features included a gaussian dip or bump along the potential, or even monodromy potentials. The best fit we found (relatively to the power-law baseline V_alpha) are depicted here:
November 4, 2025 at 9:36 AM
Such localised features included a gaussian dip or bump along the potential, or even monodromy potentials. The best fit we found (relatively to the power-law baseline V_alpha) are depicted here:
To do so, we constructed an emulator (PIPE) that is able to map the our MCMC priors for the power spectrum onto a likelihood for inflation potential parameters. We used this new tool to obtain best fit models including localised features on top of a power-law baseline potential.
November 4, 2025 at 9:36 AM
To do so, we constructed an emulator (PIPE) that is able to map the our MCMC priors for the power spectrum onto a likelihood for inflation potential parameters. We used this new tool to obtain best fit models including localised features on top of a power-law baseline potential.