Scientific seminar "Nonpotential dynamical systems and neural network technologies"

Scientific seminar "Nonpotential dynamical systems and neural network technologies"

The event passed
24 May
Location
Online
About the event

On May 24 at 19:00 p.m. (Moscow time)

Speaker: Shorokhov Sergey, associate professor of Department of Information Technology.

Topic: “Deep learning methods for solving boundary value problems with variational formulations”.

Annotation

We study approaches based on deep neural networks for solving boundary value problems for partial differential equations. Application of classical and non-classical variational principles in training neural network models that approximate solutions to boundary value problems is discussed. Examples of constructing neural network approximations for solving a boundary value problem for a second-order partial differential equation are presented.

Online

Related eventsAll events
2023
13 - 15 Dec
Seminar “Green infrastructure and sustainable development”
2023
6 - 8 Dec
Seminar “Life phase in urban soils and sediments”
2023
12 Dec
Seminar “Investigating the Placement of a MEC Application with Dynamic Cost”
2023
28 Nov - 13 Dec
The competition for the establishment of laboratories (engineering centers) under the leadership of young scientists for the implementation of scientific projects