StoProcGFlows

Материал из Wiki - Факультет компьютерных наук
Перейти к: навигация, поиск

Stochastic Processes, Geometry, Flows

Welcome! We are the new open and question-friendly seminar about mathematics called "Sto(chastic )Proc(esses, )G(eometry, )Flows". Our seminar is the place to share the knowledge, discuss peculiar stuff in mathematics and learn from each other.

It is not (fully) a research seminar, since we discuss not only the papers, cutting-edge research, but also some classic topics which are just useful for our everyday work. Why not to discuss the ideas with others if you study something? Do not forget though about real research talks which are also a part of the seminar.

It is not (fully) an educational seminar (like a course or something), since we are not aimed at very classic subjects of first 2-3 years of (applied) maths. Evidently, no one needs yet another course on Linear Algebra or Calculus 2. Instead, we are discussing the concepts which may be already widely known and yet are quite far from "university classics engraved in textbooks". The title also depicts the main interests of this seminar: Stochastic Processes, Geometry, Flows and Dynamical Systems, Functional Calculus and Operator Theory, Optimal Transport and Statistics.

So, give us a topic you want to know about or just hang out nearby for interesting discussions. We greatly encourage you to give a talk and will provide you with all necessary help and equipment. If you would like to deliver a talk, see Contacts for the details. Anyway, we'll gladly see you in the audience or near the whiteboard.

Research Areas

Stochastic Processes, Geometry, Dynamical Systems, Functional Analysis, SDE, Optimal Transport, Statistics

Resources

The community is managed through the discord, where we have various chats, archives, streams and links. To get the invite, write to maxkaledin@gmail.com.

We also (will)have a YouTube page with the archived streams and events (HREF).

Schedule, Time and Place

The time is Friday, 1900-2100 (Moscow time), the place is subject to change, check the schedule of the talks. Generally, we meet off-line with a streaming available, the archives can be also found on YouTube (HREF). The talks take no more than two hours usually, either with a board, or a presentation. Sometimes we may have a speaker online, if you are one, you are greatly welcome as well. We also recommend speakers to send some materials like abstract, drafts and notes before the day so that we could discuss the content, if you would like to have some feedback. Working language by default is English due to the reasons of going global but speakers are free to set the things otherwise if desired.

Contacts

Use the discord, it is a way full of fun to make business and have discussions:) In case you are new here, you might be interested in other contacts.

Max Kaledin (tech stuff, invites, talk proposals): maxkaledin@gmail.com, TG @XuMuK_MK

Daria Demidova (invites, talk proposals, archives): demidova.math@gmail.com, TG @dashademidova

Archive of Talks

September 24, 2021, 1900-2100(Moscow). Theory of distributions. Speaker: Max Kaledin (HDI Lab). Place: HSE (TBA) + discord.


What do you think of when you encounter the word "distribution"? In probability it is quite clear: we have random variables and their probability laws. It turns out that in western mathematical tradition this word also depicts what in Russian is known as "generalized functions". What is even more interesting: these are after a fashion the same things! Name a Dirac's delta-function as the first example producing suspicions. These objects have their own calculus (theory of distributions) which helps to generalize the calculus over the common functions in a very natural way. Distributions provide not only a new look on classic calculus problems, but also a basic tool to discuss dynamics of measures in theory of gradient flows and optimal transport.

This series is thought to be 2-week long and requires only fundamentals (Calculus 1 and 2, Linear Algebra, ODEs, some probability). A good topic for a start and a promising thing to have while sailing ⛵ in optimal transport and gradient flows.

We start with the transport equation and the definition of distributions. Then we define the operations and the fundamentals of the calculus of distributions: convergence, derivatives, multiplication by C^inf function, convolutions and jump formula (formule des sauts) in dimension 1. Finally, we will define the solution of a PDE in the sense of distributions and see what new insights we may get from it.