Abstract:We analyze the distribution of eigenvectors for mesoscopic, mean-field perturbations of diagonal matrices in the bulk of the spectrum. Our results apply to a generalized $N\times N$ Rosenzweig-Porter model. We prove that the eigenvectors entries are asymptotically Gaussian with a specific variance, localizing them onto a small, explicit part of the spectrum. For a well spread initial spectrum, this variance profile universally follows a heavy-tailed Cauchy distribution. In the case of smooth entries, we also obtain a strong form of quantum unique ergodicity as an overwhelming probability bound on the eigenvectors probability mass. The proof relies on a priori local laws for this model and the eigenvector moment flow.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is about the eigenvector distribution of deformed Wigner matrices and their Quantum Unique Ergodicity (QUE). Specifically, the author analyzes the eigenvector distribution of diagonal matrices after mesoscopic and mean - field perturbations in the middle part of the spectrum. The main research objectives include:
1. **Eigenvector Distribution**:
- It is proved that the components of eigenvectors are asymptotically Gaussian, and a specific variance expression is given.
- For the case of a well - behaved initial spectral distribution, this variance profile exhibits a heavy - tailed Cauchy distribution.
2. **Quantum Unique Ergodicity**:
- In the case of smooth components, a strong form of quantum unique ergodicity is obtained, which is manifested as a strong concentration inequality of the mass of eigenvectors on a given set of coordinates.
### Specific Problem Description
#### Eigenvector Distribution
The author considers a generalized Rosenzweig - Porter model, which performs a mesoscopic perturbation on a deterministic diagonal matrix through a Wigner random matrix. The main results include:
- **Gaussian Distribution**: The components of eigenvectors are asymptotically Gaussian in the middle part of the spectrum, and their variance depends on the initial potential, the parameter \( t \), and the position of the spectrum.
- **Cauchy Distribution**: For the case of a well - behaved initial spectral distribution, the variance profile exhibits a heavy - tailed Cauchy distribution.
#### Quantum Unique Ergodicity
- **Weak Form of Quantum Unique Ergodicity**: For any set \( I \) and eigenvector \( u_k \) within the spectral window, the deviation between the mass of the eigenvector on the set \( I \) and its expected value is small with high probability.
- **Strong Form of Quantum Unique Ergodicity**: In the case of smooth components, the strong concentration inequality of the mass of eigenvectors on a given set of coordinates shows that the mass of eigenvectors is asymptotically flatly distributed and concentrated on an explicit deterministic profile with a quantitative error.
### Method Overview
The method of the paper is based on the following three steps:
1. **Local Law**: First, through the local control of the matrix ensemble, a local law about the eigenvector projection is obtained.
2. **Short - time Relaxation**: By introducing a small Gaussian perturbation, the evolution of eigenvectors in a short time is studied.
3. **Local Statistical Invariance**: Finally, by removing the Gaussian part, the results of the original matrix \( W_t \) are obtained.
### Mathematical Formulas
- **Eigenvector Variance**:
\[
\sigma_t^2(q, k, \eta) = N \sum_{\alpha = 1}^N \frac{q_\alpha^2 t}{(D_\alpha - \gamma_{k,i,t} - t \operatorname{Re} m_t(\gamma_{k,i,t} + i\eta))^2+(t \operatorname{Im} m_t(\gamma_{k,i,t} + i\eta))^2}
\]
where \( \gamma_{k,i,t} \) is the quantile of the spectral density, and \( m_t(z) \) is the Stieltjes transform of the spectral density.
- **Weak Form of Quantum Unique Ergodicity**:
\[
P\left( \left| \frac{Nt}{|I|} \left| \sum_{\alpha \in I} |u_k(\alpha)|^2 - \frac{1}{N} \sum_{\alpha \in I} \sigma_t^2(e_\alpha, k) \right| > c \right| \right) \leq C (N^{-\vartheta}+|I|^{-1})
\]
These results provide an important theoretical basis for understanding the properties of eigenvectors in complex quantum systems.