Physics Informed Neural Networks (PINNs): "PyTorch" Solve Physical Systems with Deep Neural Networks

[@REHIMIF]

I would like to thank you so much for your efforts, GOD bless you ♥♥♥

[@Dr.Mohammad_Samara]

Thank God bless you too. 🌺

[@azeemishaq8240]

thank you so much for providing this useful information

[@Dr.Mohammad_Samara]

Glad it was helpful!

[@sarthakv1030]

Thank you!!

[@Dr.Mohammad_Samara]

Thank you for watching

[@pradipta.]

Wonderful Video 2:56
Need more informative content on it... ❤

[@Dr.Mohammad_Samara]

Thank you for your kind comment I will make more PINNs videos.

[@Rainrain0401]

so great! could you tell me the python and pytorch version

[@Dr.Mohammad_Samara]

Please watch the video we have about matching pytorch and python to make it work.

[@patelpavan5479]

make more video on pinns

[@Dr.Mohammad_Samara]

Thank you for your comment, more engineering/ PINNS will come... for now
Please see the following Video about solving PINNS using DeepXDE Library :
ruclips.net/video/L31aueuRCF8/видео.html
Also you can see the following as an intro about the topic:
ruclips.net/video/EfnSNz1egbw/видео.html

[@actuarialscience2283]

This code is not running

[@Dr.Mohammad_Samara]

Please be carful about the libraries installation and if they are compatible to your Python versions; search the error you are getting.

[@actuarialscience2283]

​@@Dr.Mohammad_SamaraThanks prof, I will do exactly that

[@azeemishaq8240]

i want to ask can we solve system of ode just like system of 5 or 6 odes coupled together using pinns technique

[@Dr.Mohammad_Samara]

Yes you can do such thing just calculate the losses of these equations in one sum of loss that need to be minimized by the Neural network.
Note that you need to equal the equations to zero and equal the right side of them to the loss value of that equation.

[@azeemishaq8240]

okay thank you so much sir@@Dr.Mohammad_Samara

[@antoninigiulio]

Is there any clear measure of the advantage of using pinns instead of more standard methods for solving pdes?

[@antoninigiulio]

Thank you

[@Dr.Mohammad_Samara]

​@@antoninigiulio PINNs can Generalize the solution that is training a network to get an answer very fast in time if you like you can check DeepONets and Inverse_PINNs videos it can show you how PINNs can be used in more practical ways.

[@fadoobaba]

how do we calculate change in physics loss wrt w1 w2 b1 b2 etc?

[@Dr.Mohammad_Samara]

Sorry I didn't exactly understand your question but for loss it need to minimize via optimization algorithms.

[@user-gz5gn9fb5h]

LGTM🍛 by Hisashi Nakai

[@Dr.Mohammad_Samara]

Nakai sann, isashiburi, welcome to foodie channel 🍗🍖🌯🍛