What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to estimate the convergence rate of the total variation distance between probability distributions and apply these estimates to multiple fields, including Malliavin calculus, normal approximation, and stochastic dynamical systems with memory. Specifically, the author proposes a method based on Fourier analysis to convert metric results related to weak convergence into estimates of the total variation distance. ### Main Problems and Solutions 1. **Estimating the Convergence Rate of the Total Variation Distance**: - The author proposes a method based on Fourier analysis to estimate the convergence rate of random variables under the total variation distance. By introducing appropriate moment conditions and smoothness conditions, the author is able to transform the results of weak convergence into estimates of the total variation distance. - The specific formula is as follows: \[ d_{\text{TV}}(X_1, X_2) \leq G d_{\text{FM}}(X_1, X_2)^g \] where \(d_{\text{FM}}\) is the Fortet - Mourier distance, and \(G\) and \(g\) are constants depending on the dimension \(d\), parameters \(\gamma\) and \(\delta\). 2. **Application Fields**: - **Malliavin Calculus**: The convergence properties of smooth functional sequences under the total variation distance are studied. - **Normal Approximation**: The multi - dimensional Berry - Esseen theorem is re - examined, providing an almost optimal convergence rate under the total variation distance. - **Stochastic Dynamical Systems with Memory**: The geometric convergence rate of stochastic recursive sequences driven by colored noise is explored. 3. **Improved Estimation Method**: - Under more stringent moment conditions and smoothness conditions, the author further improves the estimate of the total variation distance and obtains a more accurate convergence rate. For example, in the case of satisfying certain exponential moment conditions, we can obtain: \[ d_{\text{TV}}(X_1, X_2) \leq C_k d_k(X_1, X_2)^{1-\epsilon} \left| \ln d_k(X_1, X_2) \right|^{2d + k} \] ### Summary The main contribution of this paper lies in providing a new method to estimate the convergence rate of the total variation distance and applying it to several important fields in probability theory and statistics. By combining Fourier analysis and appropriate conditions, the author not only improves the existing theoretical results but also provides new directions for future research.

Rate estimates for total variation distance with applications

Convergence in Total Variation for nonlinear functionals of random hyperspherical harmonics

Metric Fourier approximation of set-valued functions of bounded variation

Normal approximation in total variation for statistics in geometric probability

A Note on the Convergence Rate of the Spectral Distributions of Large Sample Covariance Matrices

Lower bounds on the rate of convergence for accept-reject-based Markov chains in Wasserstein and total variation distances

Convergence in total variation for the kinetic Langevin algorithm

The Total Variation Distance Between the Solutions to Stochastic Volterra Equations and SDEs with Its Applications

Convergence Rate Analysis in Limit Theorems for Nonlinear Functionals of the Second Wiener Chaos

Stable central limit theorem in total variation distance

Trends to Equilibrium in Total Variation Distance

Sharp total variation rates of convergence for fluctuations of linear statistics of $β$-ensembles

New estimates for a class of non-local approximations of the total variation

A total variation version of Breuer--Major Central Limit Theorem under $\mathbb{D}^{1,2}$ assumption

Semi-Nonparametric Estimation of Distribution Divergence in Non-Euclidean Spaces

Empirical measures and random walks on compact spaces in the quadratic Wasserstein metric

Asymptotics of Distances Between Sample Covariance Matrices

Wasserstein and total variation distance between marginals of Lévy processes

On the Law of Large Numbers and Convergence Rates for the Discrete Fourier Transform of Random Fields

Learning discrete distributions with infinite support