Mathematics of AI
Follow the mathematical path from text and tokens to representations, attention and learning.
Build AI and ML intuition from tokens through attention and optimization.
- 8
- lessons
- 100
- visual explanations
- 2
- chapters
Language, Tokens & Representations
Watch human text become bytes, tokens and mathematical objects a model can process.
LESSONByte Pair Encoding: Learning Reusable Text PiecesHow can repeated neighbouring symbols become a reusable vocabulary for new text?
20 pictures · Open →
LESSONCovariance, Tensors and Attention Are Different IdeasHow do we separate a statistical relationship, a data container and an input-dependent AI operation?
13 pictures · Open →
LESSONFrom Multilingual Text to Model RepresentationsHow does one model turn many writing systems into reusable tokens and context-aware vectors?
7 pictures · Open →
LESSONStatistics Is the Foundation of Reinforcement LearningHow do repeated rewards become a value estimate and then a decision?
1 pictures · Open →
LESSONTokenization: From Human Words to Model PiecesHow should a language model divide unlimited text into a finite, reusable vocabulary?
20 pictures · Open →
LESSONTokenization: From Text to Model InputHow does ordinary text become numerical input that a language model can process?
6 pictures · Open →
LESSONUTF-8: From Unicode Code Points to BytesHow does text become an unambiguous stream of bytes before tokenization?
15 pictures · Open →Learning & Optimization Foundations
Connect loss, feedback, covariance and optimization to how learning systems improve.
