# tum-pbs/pbdl-book

Welcome to the Physics-based Deep Learning Book v0.3 - the GenAI Edition

Repository: https://github.com/tum-pbs/pbdl-book
Canonical: https://ross.abutalabs.com/products/pbdl-book
Homepage: https://physicsbaseddeeplearning.org/
Language: Jupyter Notebook
License Family: other
Topics: deep-learning, numerical-simulations, pde-solvers, artificial-intelligence, fluids, probabilistic-models, bayesian-inference, machine-learning, spatio-temporal-prediction
Last push: 2025-08-12T08:58:59+00:00

## Health v2 (maintenance only)
Score: 48/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 36, release rhythm 35, longevity 100
- inputs: {"age_days": 2067, "days_push": 386, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1361, forks 214 (observed 2026-08-28T04:04:30.024958+00:00)

## What it is
An open-source Jupyter book on combining deep learning with physical simulations and numerical methods, readable online or as a PDF. It includes interactive notebooks covering physics-informed losses, differentiable simulations, diffusion-based generative models, and reinforcement learning for PDE problems.

## Use cases
- learn how to apply deep learning to solve PDE problems
- train neural networks to predict fluid flow around airfoils
- understand differentiable simulations and physics-informed training
- build probabilistic surrogate models with diffusion models
- learn to combine neural networks with numerical simulators for inverse problems
- find a textbook on physics-based deep learning with hands-on notebooks

## When to choose
- you want a practical, notebook-driven introduction to ML for physical simulations
- you need to learn physics-informed neural networks, differentiable simulation, or diffusion-based surrogate modeling
- you have some deep learning background and want to apply it to scientific computing

## When to avoid
- you need a beginner introduction to deep learning or numerical simulation fundamentals
- you want a comprehensive survey of research papers in the field
- you need production-ready software rather than educational material

## Facets
- artifact type: learning-resource
- maturity: active
- function: machine-learning, deep-learning, simulation, nlp
- domain: machine-learning, deep-learning, simulation, artificial-intelligence, tutorials
- platform: python, cross-platform
- tags: jupyter-book, physics-based-deep-learning, pde-solvers, differentiable-simulation, diffusion-models, scientific-computing, fluids, generative-ai

## Member repositories
- tum-pbs/pbdl-book (main) score 48

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:30.024958+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:41:36.780223+00:00, confidence not recorded.
  - readme: https://github.com/tum-pbs/pbdl-book (fetched 2026-08-28T04:04:30.024958+00:00, sha 060aec5357ec)
  - homepage: https://physicsbaseddeeplearning.org/ (fetched 2026-08-29T11:59:22.473554+00:00, sha 44136fa355b3)
- Data as of 2026-08-30T08:39:29.467469+00:00.