How DeepMind learns physics simulators with Graph Networks (w/ author interview)

Join my FREE course Basics of Graph Neural Networks (www.graphneuralnets.com/p/basics-of-gnns/?src=yt)! This video dives into the paper "Learning to Simulate Complex Physics with Graph Networks" from DeepMind and interviews one of its authors, Jonathan Godwin.

Original Paper: arxiv.org/abs/2002.09405
Simulator video source: sites.google.com/view/learning-to-simulate/
Project Code & Datasets: github.com/deepmind/deepmind-research/tree/master/…

Mailing List: blog.zakjost.com/subscribe
Discord Server: discord.gg/xh2chKX
Blog: blog.zakjost.com/
Patreon: www.patreon.com/welcomeaioverlords

References:
Daniel Holden's talk from UbiSoft:    • GDC 2020 - Machine Learning, Physics ...  
SPlisHSPlasH project: github.com/InteractiveComputerGraphics/SPlisHSPlas…
"Data-driven Fluid Simulations using Regression Forests": people.inf.ethz.ch/ladickyl/fluid_sigasia15.pdf
"Latent Space Physics: Towards Learning the Temporal Evolution of Fluid Flow": arxiv.org/pdf/1802.10123.pdf
"Learning to Predict the Cosmological Structure Formation": arxiv.org/pdf/1811.06533.pdf
"Graph Networks as Learnable Physics Engines for Inference and Control": arxiv.org/pdf/1806.01242.pdf
"Relational inductive biases, deep learning, and graph networks": arxiv.org/pdf/1806.01261.pdf


Chapters
00:00 - Intro
02:24 - Why learnable physics engines?
03:15 - Literature survey
05:51 - High level overview of learning process
09:04 - Understanding the role of Graph Networks
13:15 - Interview with Jonathan Godwin introduction
14:26 - What are the key contributions of this paper?
16:40 - Why does this generalize so well?
18:23 - What about the "butterfly effect"?
21:08 - Possible application areas
25:35 - What framework for implementing/scaling this?
28:47 - Open questions and challenges
32:35 - What other research areas excite you, outside of GNNs?