arXiv subjects

Shohruh Miryusupov

Publications and source records attributed to Shohruh Miryusupov.

5 recordsLinked to original sources

arXiv2025

Reproducing kernel methods for machine learning, PDEs, and statistics

This monograph develops a unified, application-driven framework for kernel methods grounded in reproducing kernel Hilbert spaces (RKHS) and optimal transport (OT). Part I lays the theoretical and numerical foundations on positive-definite kernels; discrete and continuous RKHS; kernel engineering and scaling maps; error assessment via kernel discrepancy/maximum mean discrepancy (MMD); and a systematic operator view of kernels. In this viewpoint, projection, gradient, divergence, and Laplace-Beltrami operators are built directly from kernels, enabling discrete analogues of differential operators and variational tools that connect learning with PDE-style modeling. Part II turns to practice across four domains. In machine learning, we treat supervised and unsupervised tasks, then develop RKHS-based generative modeling, contrasting density and projection approaches and enhancing them with OT and scalable, combinatorial assignments. We introduce clustering strategies that reduce computational burden and support large-scale regression and transport. In physics-informed modeling, we present mesh-free kernel discretizations for elliptic and time-dependent PDEs, discuss automatic differentiation, and propose high-order discrete approximations. In reinforcement learning, we formulate kernel Q-learning and non-parametric HJB methods, and show how kernel operators yield sample-efficient baselines on continuous-state, discrete-action tasks. In mathematical finance, we build nonparametric time-series models and market generators, study benchmarking and extrapolation for pricing, and apply the framework to stress testing and portfolio methods.

math.NA↗

arXiv2024

A class of kernel-based scalable algorithms for data science

We present several generative and predictive algorithms based on the RKHS (reproducing kernel Hilbert spaces) methodology, which, most importantly, are scale up efficiently with large datasets or high-dimensional data. It is well recognized that the RKHS methodology leads one to efficient and robust algorithms for numerous tasks in data science, statistics, and scientific computation. However, the implementations existing the literature are often difficult to scale up for encompassing large datasets. In this paper, we introduce a simple and robust, divide-and-conquer methodology. It applies to large scale datasets and relies on several kernel-based algorithms, which distinguish between various extrapolation, interpolation, and optimal transport steps. We argue how to select the suitable algorithm in specific applications thanks to a feedback of performance criteria. Our primary focus is on applications and problems arising in industrial contexts, such as generating meshes for efficient numerical simulations, designing generators for conditional distributions, constructing transition probability matrices for statistical or stochastic applications, and addressing various tasks relevant to the Artificial Intelligence community. The proposed algorithms are highly relevant to supervised and unsupervised learning, generative methods, as well as reinforcement learning.

math.NA↗

arXiv2024

Extrapolation and generative algorithms for three applications in finance

For three applications of central interest in finance, we demonstrate the relevance of numerical algorithms based on reproducing kernel Hilbert space (RKHS) techniques. Three use cases are investigated. First, we show that extrapolating from few pricer examples leads to sufficiently accurate and computationally efficient results so that our algorithm can serve as a pricing framework. The second use case concerns reverse stress testing, which is formulated as an inversion function problem and is treated here via an optimal transport technique in combination with the notions of kernel-based encoders, decoders, and generators. Third, we show that standard techniques for time series analysis can be enhanced by using the proposed generative algorithms. Namely, we use our algorithm in order to extend the validity of any given quantitative model. Our approach allows for conditional analysis as well as for escaping the `Gaussian world'. This latter property is illustrated here with a portfolio investment strategy.

math.NA↗

arXiv2017

Optimal Transport Filtering with Particle Reweighing in Finance

We show the application of an optimal transportation approach to estimate stochastic volatility process by using the flow that optimally transports the set of particles from the prior to a posterior distribution. We also show how to direct the flow to a rarely visited areas of the state space by using a particle method (a mutation and a reweighing mechanism). We demonstrate the efficiency of our approach on a simple example of the European option price under the Stein-Stein stochastic volatility model for which a closed form formula is available. Both homotopy and reweighted homotopy methods show a lower variance, root-mean squared errors and a bias compared to other filtering schemes recently developed in the signal-processing literature, including particle filter techniques.

math.NA↗

arXiv2017

Hamiltonian Flow Simulation of Rare Events

Hamiltonian Flow Monte Carlo(HFMC) methods have been implemented in engineering, biology and chemistry. HFMC makes large gradient based steps to rapidly explore the state space. The application of the Hamiltonian dynamics allows to estimate rare events and sample from target distributions defined as the change of measures. The estimates demonstrated a variance reduction of the presented algorithm and its efficiency with respect to a standard Monte Carlo and interacting particle based system(IPS). We tested the algorithm on the case of the barrier option pricing.

stat.CO↗