Abstract:Hierarchical decays of $N$ matter species to radiation may balance against Hubble expansion to yield stasis, a new phase of cosmological evolution with constant matter and radiation abundances. We analyze stasis with various machine learning techniques on the full $2N$-dimensional space of decay rates and abundances, which serve as inputs to the system of Boltzmann equations that governs the dynamics. We construct a differentiable Boltzmann solver to maximize the number of stasis $e$-folds $\mathcal{N}$. High-stasis configurations obtained by gradient ascent motivate log-uniform distributions on rates and abundances to accompany power-law distributions of previous works. We demonstrate that random configurations drawn from these families of distributions regularly exhibit many $e$-folds of stasis. We additionally use them as priors in a Bayesian analysis conditioned on stasis, using stochastic variational inference with normalizing flows to model the posterior. All three numerical analyses demonstrate the generality of stasis and point to a new model in which the rates and abundances are exponential in the species index. We show that the exponential model solves the exact stasis equations, is an attractor, and satisfies $\mathcal{N}\propto N$, exhibiting inflation-level $e$-folding with a relatively low number of species. This is contrasted with the $\mathcal{N}\propto \log(N)$ scaling of power-law models. Finally, we discuss implications for the emergent string conjecture and string axiverse.
Cosmology and Nongalactic Astrophysics,High Energy Physics - Phenomenology,High Energy Physics - Theory
What problem does this paper attempt to address?
The problem that this paper attempts to solve is whether, in cosmology, during the expansion of the universe, the abundances of matter, radiation, and/or vacuum energy can remain unchanged over a long period of time, that is, whether the universe can experience a "stasis" phase. Specifically, the author explores the possibility of a new cosmic evolution phase in which the abundances of matter and radiation are constant by balancing Hubble expansion through hierarchical decays between different species. The focus of the paper is on analyzing the generality and persistence of this stasis state and proposes a method of using machine - learning techniques to optimize the stasis state.
### Background of the Paper and Problem Definition
1. **Temporal Evolution in Cosmology**
- One of the标志性 characteristics of the expansion of the universe is the change over time of each component that makes up the total energy density of the universe. However, recent research has proposed that this assumption does not always hold, and the universe may experience a long - term "stasis" phase during which the abundances of matter, radiation, and/or vacuum energy in the universe can remain stable.
- This stasis phenomenon can be achieved by energy pumps driven by various physical mechanisms, which can counteract the effects of the expansion of the universe.
2. **Different Types of Stasis**
- In the original definition, stasis is achieved through the hierarchical decay of a series of matter states \(\phi_\ell\) into radiation (referred to as \(M\rightarrow\gamma\) stasis).
- Other studies have also introduced stasis mechanisms for the conversion of vacuum energy into matter (\(\Lambda\rightarrow M\)) and radiation (\(\Lambda\rightarrow\gamma\)), as well as "triple stasis" that includes all three conversions simultaneously.
3. **Physical Theoretical Basis**
- These stasis phenomena can naturally arise from various physical theories beyond the Standard Model (BSM), such as theories that predict the existence of unstable particle towers.
### Research Methods and Objectives
1. **Numerical Analysis and Machine Learning**
- The author uses machine - learning techniques to analyze the decay rates and abundances in the entire 2N - dimensional space, with these parameters as inputs to the Boltzmann equations that control the dynamics.
- A differentiable Boltzmann solver is constructed to maximize the number of e - folds \(N\) in the stasis state.
- High - stasis configurations are obtained through the gradient - ascent method, which inspire a log - uniform distribution of decay rates and abundances, supplementing the previous power - law distribution.
2. **Random Configuration and Bayesian Analysis**
- Randomly sampling configurations in these distribution families usually shows many e - folds of stasis.
- Using these distributions as prior conditions, Bayesian analysis is carried out, and the posterior distribution is modeled by a normalizing flow model through stochastic variational inference.
3. **Discovery of the Exponential Model**
- Numerical analysis reveals a new exponential model that can solve the exact stasis equations, is an attractor, and satisfies the relationship \(N\propto N\), exhibiting a number of e - folds close to the expansion level, while the required number of species is relatively small.
- Compared with the \(N\propto\log(N)\) relationship of the power - law model, the exponential model shows a longer stasis period.
### Conclusions and Significance
1. **Generality of Stasis**
- All three numerical analyses indicate that the stasis state is widespread and point to a new model in which the decay rates and abundances change exponentially with the species index.
- The exponential model can not only explain the long - term stasis phenomenon but may also provide a new perspective on certain conjectures in string theory (such as axiverse).
2. **Physical and Cosmological Significance**
- This stasis phenomenon has important implications for the early evolution and observation of the universe, especially in the period after nucleosynthesis, and may impose constraints on ultraviolet - complete theories (such as string theory).
In summary, through theoretical analysis and numerical simulation, this paper explores the universality and persistence of the stasis state of the universe and proposes a new exponential model, providing a new perspective for understanding the complex dynamics in the early universe.