Drift Estimation of Multiscale Diffusions Based on Filtered Data

We study the problem of drift estimation for two-scale continuous time series. We set ourselves in the framework of overdamped Langevin equations, for which a single-scale surrogate homogenized equation exists. In this setting, estimating the drift coefficient of the homogenized equation requires pre-processing of the data, often in the form of subsampling; this is because the two-scale equation and the homogenized single-scale equation are incompatible at small scales, generating mutually singular measures on the path space. We avoid subsampling and work instead with filtered data, found by application of an appropriate kernel function, and compute maximum likelihood estimators based on the filtered process. We show that the estimators we propose are asymptotically unbiased and demonstrate numerically the advantages of our method with respect to subsampling. Finally, we show how our filtered data methodology can be combined with Bayesian techniques and provide a full uncertainty quantification of the inference procedure.

Restriction of Exponential Sums to Hypersurfaces

We prove moment inequalities for exponential sums with respect to singular measures, whose Fourier decay matches those of curved hypersurfaces. Our emphasis will be on proving estimates that are sharp with respect to the scale parameter $N$, apart from $N^\epsilon $ losses. In a few instances, we manage to remove these losses.

The existence of minimizers for an isoperimetric problem with Wasserstein penalty term in unbounded domains

In this article, we consider the (double) minimization problem min ⁡ { P ⁢ ( E ; Ω ) + λ ⁢ W p ⁢ ( E , F ) : E ⊆ Ω , F ⊆ R d , | E ∩ F | = 0 , | E | = | F | = 1 } , \min\{P(E;\Omega)+\lambda W_{p}(E,F):E\subseteq\Omega,\,F\subseteq\mathbb{R}^{d},\,\lvert E\cap F\rvert=0,\,\lvert E\rvert=\lvert F\rvert=1\}, where λ ⩾ 0 \lambda\geqslant 0 , p ⩾ 1 p\geqslant 1 , Ω is a (possibly unbounded) domain in R d \mathbb{R}^{d} , P ⁢ ( E ; Ω ) P(E;\Omega) denotes the relative perimeter of 𝐸 in Ω and W p W_{p} denotes the 𝑝-Wasserstein distance. When Ω is unbounded and d ⩾ 3 d\geqslant 3 , it is an open problem proposed by Buttazzo, Carlier and Laborde in the paper On the Wasserstein distance between mutually singular measures. We prove the existence of minimizers to this problem when the dimension d ⩾ 1 d\geqslant 1 , 1 p + 2 d > 1 \frac{1}{p}+\frac{2}{d}>1 , Ω = R d \Omega=\mathbb{R}^{d} and 𝜆 is sufficiently small.

Generalized powers and measures

Using the winding of measures on torus in "rational directions" special classes of unitary operators and pairs of isometries are defined. This provides nontrivial examples of generalized powers. Operators related to winding Szegö-singular measures are shown to have specific properties of their invariant subspaces.