Universal Differential Equations for Scientific Machine Learning - Chris Rackauckas MIT

[@AJ-et3vf]

Great video. Thank you.

[@kallolroy5029]

Great idea Chris

[@kiranshila]

Another awesome video! Thanks Chris!

[@orrymr]

Thanks for this talk :). Are the slides publicly available?

[@chrisrackauckasofficial]

Yes, a version of the slides are up at figshare.com/articles/presentation/Universal_Differential_Equations_for_Scientific_Machine_Learning/12751937

[@youngjin8300]

I've been informally putting my thoughts into this matter, or something along this line. I would love to see more researches related to this. I checked your paper on arxiv. Is there more papers or books you could refer me to?

[@chrisrackauckasofficial]

I gathered an explanation of these kinds of methods in the MIT 18.337 lecture notes: github.com/mitmath/18337 . These reference a lot of different primary literature throughout as well. However, I don't know of a book to refer to. If there's any book that's useful, it would be Griewank's Automatic Differentiation tome, but indeed that doesn't cover the scientific aspects like PDEs.

[@chandrasekharjinendran6844]

Thanks Christopher

[@andres_pq]

Hello Chris, wild idea here. What if this procedure can be used to extend the SIR model for pandemics? Maybe learning future parameters for better modeling the next illness-19. Please contact me if you are interested, I'm currently studying my masters' degree.

[@chrisrackauckasofficial]

See covid19ml.org/ and www.medrxiv.org/content/10.1101/2020.04.03.20052084v1