Conveners
Probability, Statistics and Financial Mathematics
- Zoran Vondracek (University of Zagreb)
Random Projections have been widely used to generate embeddings for various large graph tasks due to their computational efficiency in estimating relevance between vertices. The majority of applications have been justified through the Johnson-Lindenstrauss Lemma. We take a step further and investigate how well dot product and cosine similarity are preserved by Random Projections. Our analysis...
In case of an equidistant sampling of an ergodic diffusion path such that
the maximal time of observation tends to infinity and size of time interval subdivision tends to zero, we have investigated an asymptotic normality of the difference between approximate maximum likelihood estimator (AMLE) and maximum likelihood estimator based on continuous observation (MLE).
We use this property for...
The concept of regular variation plays pivotal role in understanding the extreme behavior of stochastic processes. It also finds applications in various areas ranging from random networks to stochastic geometry. We discuss some developments in this field and show how one can generalize regular variation to rather abstract settings, provided compatible notions of scaling and boundedness can be...
In this talk I will discuss the potential theory of Dirichlet forms on the half-space $\mathbb{R}^d_+$ defined by the jump kernel $J(x,y)=|x-y|^{-d-\alpha}\mathcal{B}(x,y)$ and the killing potential $\kappa x_d^{-\alpha}$, where $\alpha\in (0, 2)$ and $\mathcal{B}(x,y)$ can blow up to infinity at the boundary. The jump kernel and the killing potential depend on several parameters. For all...
