How DeepMind learns physics simulators with Graph Networks (w/ author interview)
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: https://arxiv.org/abs/2002.09405
Simulator video source: https://sites.google.com/view/learning-to-simulate/
Project Code & Datasets: https://github.com/deepmind/deepmind-research/tree/master/learning_to_simulate
Mailing List: https://blog.zakjost.com/subscribe
Discord Server: https://discord.gg/xh2chKX
Blog: https://blog.zakjost.com
Patreon: https://www.patreon.com/welcomeaioverlords
References:
- Daniel Holden's talk from UbiSoft: https://www.youtube.com/watch?v=sUb0W5_waRI
- SPlisHSPlasH project: https://github.com/InteractiveComputerGraphics/SPlisHSPlasH
- "Data-driven Fluid Simulations using Regression Forests": https://people.inf.ethz.ch/ladickyl/fluid_sigasia15.pdf
- "Latent Space Physics: Towards Learning the Temporal Evolution of Fluid Flow": https://arxiv.org/pdf/1802.10123.pdf
- "Learning to Predict the Cosmological Structure Formation": https://arxiv.org/pdf/1811.06533.pdf
- "Graph Networks as Learnable Physics Engines for Inference and Control": https://arxiv.org/pdf/1806.01242.pdf
- "Relational inductive biases, deep learning, and graph networks": https://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?
Видео How DeepMind learns physics simulators with Graph Networks (w/ author interview) канала WelcomeAIOverlords
Original Paper: https://arxiv.org/abs/2002.09405
Simulator video source: https://sites.google.com/view/learning-to-simulate/
Project Code & Datasets: https://github.com/deepmind/deepmind-research/tree/master/learning_to_simulate
Mailing List: https://blog.zakjost.com/subscribe
Discord Server: https://discord.gg/xh2chKX
Blog: https://blog.zakjost.com
Patreon: https://www.patreon.com/welcomeaioverlords
References:
- Daniel Holden's talk from UbiSoft: https://www.youtube.com/watch?v=sUb0W5_waRI
- SPlisHSPlasH project: https://github.com/InteractiveComputerGraphics/SPlisHSPlasH
- "Data-driven Fluid Simulations using Regression Forests": https://people.inf.ethz.ch/ladickyl/fluid_sigasia15.pdf
- "Latent Space Physics: Towards Learning the Temporal Evolution of Fluid Flow": https://arxiv.org/pdf/1802.10123.pdf
- "Learning to Predict the Cosmological Structure Formation": https://arxiv.org/pdf/1811.06533.pdf
- "Graph Networks as Learnable Physics Engines for Inference and Control": https://arxiv.org/pdf/1806.01242.pdf
- "Relational inductive biases, deep learning, and graph networks": https://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?
Видео How DeepMind learns physics simulators with Graph Networks (w/ author interview) канала WelcomeAIOverlords
Показать
Комментарии отсутствуют
Информация о видео
Другие видео канала
Graph Convolutional Networks (GCNs) made simpleQuantum Computers, Explained With Quantum PhysicsAlphaFold: The making of a scientific breakthroughHow Well Can an AI Learn Physics? ⚛Magnetohydrodynamics - Propelling Liquid Metal with Magnets!DeepMind’s Demis Hassabis on its breakthrough scientific discoveries | WIRED LiveIntro to Relational - Graph Convolutional NetworksHow Are Highways Designed?Intro to Graphs and Label Propagation Algorithm in Machine LearningInterview w/ Former Neuralink Engineer: Han ZhangCombatting Anti-Science with Richard DawkinsDr. Michio Kaku Answers Physics Questions From Twitter | Tech Support | WIREDMuZero: Mastering Atari, Go, Chess and Shogi by Planning with a Learned ModelGraph Representation Learning (Stanford university)What Is Dynamic Programming and How To Use ItSimple Explanation of AutoEncodersDDPS | Learning physical simulation with graph networkHow to use Adversarial Validation to Help Fix OverfittingDeep Recommender Systems at Facebook feat. Carole-Jean Wu | Stanford MLSys Seminar Episode 24K Jarrod Millman - Complex network analysis with NetworkX| PyData Global 2020