A Few Pieces, Endlessly Scaled
Nudge the Knob, Listen
Take a radio with a single tuning knob and an unfamiliar station you want to find. Turn the knob a little and listen: is the music clearer or noisier than a moment ago?
Library · CC BY-SA 4.0
Alice's Adventures in aDifferentiable Wonderland
Alice's Adventures in a Differentiable Wonderland, Volume I: A Tour of the Land
Choose your version
The same text, told at every level, in every language and look shown here. Pick, then listen or download.
VoiceMale · Charon
DrawingsGraphic inkWhite on black
Listen and watch
The same player the product gives for your own book: the voice, the drawings as they are drawn, the text.
A modern textbook on how neural networks really work, from tensors and gradients to transformers. The first three chapters lay the ground: what the book is about, the linear algebra and calculus a network runs on, gradient descent, and what a dataset and a loss function are.
This is a sample: 3 of 13 chapters.
From the lectures
Arrows That Point the Same Way
Cosine Similarity
The most important vector operation in the whole book is the dot product, also called the inner product. You multiply the two vectors entry by entry and add up everything you got, so the result is a single number, not a vector.
A Recipe in Letters, and the Slope of a Curve
The Slope of a Touching Line
Draw a smooth curve that dips down and then climbs again, like a shallow valley, and choose a point on its slope. Lay a ruler so that it just touches the curve at that point without cutting through it.
Walking Downhill in the Fog
Step by Step to the Bottom
Now the main event: finding the lowest point of a function, its minimum. The book assumes the function gives a single number and that the input may wander anywhere.
From Practice Code to Real Data
Similar Things Sit Close
The chapter's figure adds a picture worth keeping. A trained model transforms its input step by step, and the inner versions of the data it produces often have a meaningful geometry: images with similar meaning land at nearby points in an inner space.
Memorising Is Not Learning
Answers as Distributions
Predicting one single answer makes sense when, for each input, almost all of the probability sits on one value. The book relaxes that: instead of one answer, the model's outputs set the knobs of a whole probability distribution over the possible answers.
How it was made
- Level 4 · Ideen und Beispiele
- Demonstrator
- White on black · regular
- questions
- places in the book
- worked examples
- Voice: Charon (Gemini)
- Told in English
Examples for: a curious adult with school maths who uses AI apps every day; add one modern instance per idea, such as how a chat assistant or a phone's photo app is trained
The source
The lectures retell the English text of arXiv 2404.17625 (version 3, 2025) by Simone Scardapane, published under an open licence that allows retelling and translation. The retelling, the drawings and the voice are made by AI and checked against the book by a second model. How AI is used.
The lectures adapt the original and are released under the same licence, free to share and adapt with credit to the author and a link to this page: CC BY-SA 4.0
Your own hard book, the same way
Upload a textbook or any difficult book. The first five minutes are free.
Upload a book