arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2203.09677math.DScs.ITmath-phmath.DGmath.ITmath.MPmath.PR
收藏

Hölder平衡概率流形上的测地线与动力学信息投影

Geodesics and dynamical information projections on the manifold of Hölder equilibrium probabilities

Artur O. Lopes, Rafael O. Ruggiero

AI总结:

本文在Hölder平衡概率流形上研究两类几何问题:证明在Ruelle算子核具有类Fourier基时存在测地路径,并将KL散度最小化的信息投影问题推广到动力学不变概率集合,得到显式方程并探讨三角与Pythagorean不等式。

摘要:

我们在此考虑由变换 $T:M \to M$ 描述的离散时间动力学,其中 $T$ 要么是符号空间 $M=\{1,2,...,d\}^\mathbb{N}$ 上的移位作用 $T=\sigma$,要么 $T$ 描述一个 $d$ 到 $1$ 的 $C^{1+\alpha}$ 类扩张变换 $T:S^1 \to S^1$ 的作用(例如 $x \to T(x)=d\\,x$ (mod $1$)),此时 $M=S^1$ 是单位圆。已知 Hölder 势 $A:M \to \mathbb{R}$ 的平衡概率构成的无穷维流形 $\mathcal{N}$ 是一个解析流形,并且带有与渐近方差相关联的自然黎曼度量。我们在此证明,在 Ruelle 算子的核存在类 Fourier Hilbert 基的假设下,存在测地路径。当 $T=\sigma$ 且 $M=\{0,1\}^\mathbb{N}$ 时,这样的基存在。在另一个方向上,我们还考虑一对平衡概率的 KL 散度 $D_{KL}(\mu_1,\mu_2)$。若 $D_{KL}(\mu_1,\mu_2)=0$,则 $\mu_1=\mu_2$。尽管 $D_{KL}$ 不是 $\mathcal{N}$ 中的度量,但它描述了 $\mu_1$ 与 $\mu_2$ 之间的邻近程度。一个自然的问题是:对于固定的概率 $\mu_1\in \mathcal{N}$,考虑 $\mathcal{N}$ 中某个凸概率集合内使 $D_{KL}(\mu_1,\mu_2)$ 最小化的概率 $\mu_2$。这个最小化问题是信息投影中所考虑的主要问题的一个动力学版本。我们在 $\mathcal{N}$ 中考虑该问题,此时所有概率都是动力学不变的,并得到所求解的显式方程。文中还将研究三角不等式和 Pythagorean 不等式。

英文摘要:

We consider here the discrete time dynamics described by a transformation $T:M \to M$, where $T$ is either the action of shift $T=σ$ on the symbolic space $M=\{1,2,...,d\}^\mathbb{N}$, or, $T$ describes the action of a $d$ to $1$ expanding transformation $T:S^1 \to S^1$ of class $C^{1+α}$ (\,for example $x \to T(x) =d\, x $ (mod $1) $\,), where $M=S^1$ is the unit circle. It is known that the infinite-dimensional manifold $\mathcal{N}$ of equilibrium probabilities for Hölder potentials $A:M \to \mathbb{R}$ is an analytical manifold and carries a natural Riemannian metric associated with the asymptotic variance. We show here that under the assumption of the existence of a Fourier-like Hilbert basis for the kernel of the Ruelle operator there exists geodesics paths. When $T=σ$ and $M=\{0,1\}^\mathbb{N}$ such basis exists. In a different direction, we also consider the KL-divergence $D_{KL}(μ_1,μ_2)$ for a pair of equilibrium probabilities. If $D_{KL}(μ_1,μ_2)=0$, then $μ_1=μ_2$. Although $D_{KL}$ is not a metric in $\mathcal{N}$, it describes the proximity between $μ_1$ and $μ_2$. A natural problem is: for a fixed probability $μ_1\in \mathcal{N}$ consider the probability $μ_2$ in a convex set of probabilities in $\mathcal{N}$ which minimizes $D_{KL}(μ_1,μ_2)$. This minimization problem is a dynamical version of the main issues considered in information projections. We consider this problem in $\mathcal{N}$, a case where all probabilities are dynamically invariant, getting explicit equations for the solution sought. Triangle and Pythagorean inequalities will be investigated.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑