dmivilensky Goto Github PK

accelerated-coordinate-descent-method

The implementation of Coordinate Descent Method Accelerated by Universal Metaalgorithm with efficient amortised complexity of iteration & Experiments with sparse SoftMax function, where the proposed method is better than FGM

adaptive-relatively-smooth-optimization

We introduce the new concept of (α,L,δ)-relative smoothness (see https://arxiv.org/pdf/2107.05765.pdf) which covers both the concept of relative smoothness and relative Lipschitz continuity. For the corresponding class of problems, we propose some adaptive and universal methods which have optimal estimates of the convergence rate.

canonical-barriers

Numerical solution of Painlevé III D7 equation and calculation of canonical barriers according to arXiv:1806.06588

cayley-non-flat-embeddings

Let’s train various models to embed Cayley graphs to hyperbolic plane and then study, how much can we say about groups from these embeddings. At the same time, let’s study models generalization ability, by changing their hyperparameters.

compilers-course-jb

goidle

highlight

HighLight – platform for teamworking on translations of medical documents

lambda-algebra-tools

network-utility-maximization-2021

neural-homotopy-groups-calculation

sampling-from-homotopy-groups-of-spheres

Sampling from π_n(S^2). Application of optimization and machine learning methods to problems of algebraic topology

sirius-2021-dpd-optimization

Experiments on stochastic optimization for digital pre-distortion of the signal

sliding-for-kernel-svm

We consider a problem of minimizing a sum of two functions and propose a generic algorithmic framework (SAE) to separate oracle complexities for each function. We compare the performance of splitting accelerated enveloped accelerated variance reduced method with a different sliding technique.

solar-method-non-convex-optimisation

torch-alternating

This module provides the extension for PyTorch toolkit, containing imlplementations of some alternating optimization methods acting as envelopes for arbitrary optimizers applied to the problem considered.

typology-of-caritive

This repo is a part of a typological study of caritive — a cluster of meanings describing the absence of an object at the disposal of some participant of the situation. We study the statistical properties of expert measurements of caritive indicators to find a dependencies between them

wasserstein-distance-calculation

Optimal Transport and Optimization related experiments.