intsystems / 2024-Project-162

Название исследуемой задачи

Adaptive sampling methods using diffusion models

Тип научной работы

M1P/НИР/CoIS

Автор

Марат Айратович Хусаинов

Научный руководитель

Самсонов Сергей Владимирович

Научный консультант(при наличии)

степень, Фамилия Имя Отчество

Abstract

In this paper, we consider a novel approach to sampling from probability distributions using Generative Flow Networks (GFlowNets). GFlowNets are a class of amortized inference algorithms designed to address the challenge of generating samples proportional to given rewards in a Markov decision process (MDP). In this work, we adapt GFlowNets to the task of sampling from complex continuous probability distributions, leveraging their sequential decision-making framework to efficiently explore high-dimensional spaces. Research publications =============================== 1.

Presentations at conferences on the topic of research ================================================ 1.

Software modules developed as part of the study

1. A python package mylib with all implementation here.
2. A code with all experiment visualisation here. Can use colab.