The Tools: Probabilistic Programming with GenJAX

Teach the laptop to imagine days. Over Golden Week, Jamal lends Chibany a laptop — and Part I’s counting gets a superpower. In Foundations you learned that probability questions are really three questions: What’s possible? What am I interested in? Count them! GenJAX is a probabilistic programming language that lets the computer do all three: write code that generates possible days, filter for the ones you care about, and let simulation do the counting.

Designed for complete beginners to programming. We use Google Colab (nothing to install, runs in the browser), teach just enough Python, and connect every line of code back to the sets you already know. You won’t become a programmer — you’ll become someone who can use probabilistic programming to explore probability and build models.

graph TB
    A[Getting Started] --> B[Python Basics]
    B --> C[First Model]
    C --> D[Traces]
    D --> E[Conditioning]
    E --> F[Inference in Action]
    F --> G[Building Models]

    style C fill:#f39c12,color:#fff
    style E fill:#f39c12,color:#fff
    style F fill:#f39c12,color:#fff

Core chapters (yellow): generative models, conditioning, and inference — the three ideas every later Part leans on. Later chapters’ code assumes this Part; whenever a code block puzzles you, come back here.

Wondering why this book teaches GenJAX instead of Stan or PyMC? The honest answer.

Chapters

Learning Tip

Don’t try to memorize Python syntax! Focus on what the code is trying to do, how it maps onto the probability concepts, and what happens when you run it. You can always copy-paste and modify examples — understanding beats memorization.

This project is generously funded by the Japanese Probabilistic Computing Consortium Association (JPCCA).