Differentiable Physics (for Deep Learning), Overview Talk by Nils Thuerey

Автор: Nils Thuerey

Загружено: 2020-07-01

Просмотров: 9740

Описание: In this talk Nils explains recent research works that shows how to employ differentiable PDE solvers for deep learning. A central aim here is to improve the outcomes of numerical simulations. Results for reducing the numerical error in iterative solvers, and long-term control of physical systems are demonstrated. Source code is available at: https://github.com/tum-pbs/PhiFlow

The two research papers discussed in the talk can be found at:
https://ge.in.tum.de/publications/202... & https://arxiv.org/pdf/2007.00016
https://ge.in.tum.de/publications/202... & https://arxiv.org/pdf/2001.07457

Addendum: the missing video at 33:45 is https://ge.in.tum.de/download/2020-ic... , it shows a much more "intuitive" solution compared to the single-case optimization.

Не удается загрузить Youtube-плеер. Проверьте блокировку Youtube в вашей сети.
Повторяем попытку...

Differentiable Physics (for Deep Learning), Overview Talk by Nils Thuerey

Доступные форматы для скачивания:

Скачать видео

Информация по загрузке:

Скачать аудио

Похожие видео

Adaptive Phase-Field-FLIP for Very Large Scale Two-Phase Fluid Simulation, SIGGRAPH '25

Adaptive Phase-Field-FLIP for Very Large Scale Two-Phase Fluid Simulation, SIGGRAPH '25

Nils Thuerey - Phiflow: A differentiable PDE solving framework for deep learning

Nils Thuerey - Phiflow: A differentiable PDE solving framework for deep learning

Физически-информированные нейронные сети (PINN) [Машинное обучение с учетом физики]

Физически-информированные нейронные сети (PINN) [Машинное обучение с учетом физики]

ORBAN TRACI GRUNT, a EUROPA przyznaje się do SŁABOŚCI w MONACHIUM #BizWeek

ORBAN TRACI GRUNT, a EUROPA przyznaje się do SŁABOŚCI w MONACHIUM #BizWeek

Introducing the Physics-based Deep Learning book (PBDL)

Introducing the Physics-based Deep Learning book (PBDL)

Differentiable Physical Simulation and AI @ NeurIPS 2020 DiffCVGP workshop

Differentiable Physical Simulation and AI @ NeurIPS 2020 DiffCVGP workshop

Почему скорость света слишком медленная, чтобы добраться до других галактик | Документальный фильм

Почему скорость света слишком медленная, чтобы добраться до других галактик | Документальный фильм

Learning Robot Control: From RL to Differential Simulation - (PhD Defense of Yunlong Song)

Learning Robot Control: From RL to Differential Simulation - (PhD Defense of Yunlong Song)

What is Automatic Differentiation?

What is Automatic Differentiation?

Deep Learning to Discover Coordinates for Dynamics: Autoencoders & Physics Informed Machine Learning

Deep Learning to Discover Coordinates for Dynamics: Autoencoders & Physics Informed Machine Learning

ADL4P , Lecture 01 , Introduction

ADL4P , Lecture 01 , Introduction

Why Returning From Mars ls Impossible - Richard Feynman's Warning

Why Returning From Mars ls Impossible - Richard Feynman's Warning

Physics Informed Machine Learning: High Level Overview of AI and ML in Science and Engineering

Physics Informed Machine Learning: High Level Overview of AI and ML in Science and Engineering

An extension of Physics Informed Neural Networks with a modal approach – IVADO Octobre Numérique

An extension of Physics Informed Neural Networks with a modal approach – IVADO Octobre Numérique

Teaching Neural Network to Solve Navier-Stokes Equations

Teaching Neural Network to Solve Navier-Stokes Equations

Zongyi Li's talk on solving PDEs from data

Zongyi Li's talk on solving PDEs from data

NeurIPS 2020 Tutorial: Deep Implicit Layers

NeurIPS 2020 Tutorial: Deep Implicit Layers

DDPS | "When and why physics-informed neural networks fail to train" by Paris Perdikaris

Yuanming Hu - Physical simulation and AI: Differentiability

Yuanming Hu - Physical simulation and AI: Differentiability

"Machine Learning for Partial Differential Equations" by Michael Brenner