Physics-Based Deep Learning Book
The Physics-based Deep Learning Book (v0.2) introduces deep learning for physical simulations, covering topics like physical loss constraints, tailored training algorithms, and uncertainty modeling. It includes Jupyter notebooks for practical learning.
Read original article
The Physics-based Deep Learning Book (v0.2) provides a comprehensive introduction to deep learning in the context of physical simulations. It covers topics such as physical loss constraints, differentiable simulations, tailored training algorithms for physics problems, reinforcement learning, and uncertainty modeling. The document includes hands-on code examples in the form of Jupyter notebooks for practical learning. The new version (v0.2) introduces sections on integrating differentiable physics into neural network training and improved learning methods for physics problems. Future chapters will explore training networks to infer fluid flows, using model equations as residuals for training, and interacting with simulators for inverse problems. The book emphasizes physics-based deep learning (PBDL) approaches and discusses the integration of physical models into deep learning algorithms. The text is maintained by the Physics-based Simulation Group at TUM and welcomes feedback for continuous improvement. The project acknowledges contributions from various individuals and provides a citation for referencing the book.
Related
16 commentsThis is a book on the deep learning approaches for physics problems DEVELOPED BY THIS RESEARCH GROUP. I think that is a very very important disclaimer to this book.
In addition, it is essentially used to strongly push their simulation framework Phi-Flow.
I would NOT call this an accurate depiction of the field.
https://www.stochasticlifestyle.com/the-essential-tools-of-s...
Afaik, it's produced by Jupyter book[1], but find nothing in their docs either.
1. CRUNCH group YouTube (talks on Math + ML) - https://m.youtube.com/channel/UC2ZZB80udkRvWQ4N3a8DOKQ
2. Steve Brunton's Physics Informed Machine Learning playlist - https://m.youtube.com/playlist?list=PLMrJAkhIeNNQ0BaKuBKY43k...
3. The book "Data Driven Science and Engineering" from Steve Brunton
4. Deep Learning in Scientific Computation from ETH Zurich - https://m.youtube.com/playlist?list=PLJkYEExhe7rYY5HjpIJbgo-...
> This document contains a practical and comprehensive introduction of everything related to deep learning in the context of physical simulations. As much as possible, all topics come with hands-on code examples in the form of Jupyter notebooks to quickly get started. Beyond standard supervised learning from data, we’ll look at physical loss constraints, more tightly coupled learning algorithms with differentiable simulations, training algorithms tailored to physics problems, as well as reinforcement learning and uncertainty modeling. We live in exciting times: these methods have a huge potential to fundamentally change what computer simulations can achieve.
That said, it is a lovely set of topics.