A CONNECTED PLACE TO LEARN
One idea leads
to another.
Explore the language of mathematics, the laws of physics and the systems of biology. Then use code, learning algorithms and simulations to ask your own questions. Layerchecklist brings the maps, lessons and experiments together, one layer at a time.
It is organised in four parts. The foundations come first because they give you a language for what follows. The tools turn an idea into something that runs. Machine learning explores how systems learn from data. The simulation lab lets you experiment with models of the world. Follow the path, or start with a question that interests you.
PART ONE
The foundations
Begin with Math, continue into Physics, then explore Biology. Each subject opens with an interactive atlas of selected ideas, arranged by branch and study level. Follow a connection and see where it takes you.
- MathFrom what a number is to the open problems, by way of linear algebra, calculus and probability.265 ideas mapped
- PhysicsMatter, motion, fields and chance, with connections to living systems and computing.151 ideas mapped
- BiologyLiving systems: cells, heredity, evolution and ecology, with interactive models.112 ideas mapped
- Reading the notationTranslate θ, ∇, Σ and argmin into plain sentences. Start here if equations are the obstacle.Symbol by symbol
PART TWO
The tools
Turn an idea into something you can run. Write a program, work with arrays and data, explore parallel computation, and connect the parts of a robot.
- Python FundamentalsTypes, control flow, functions and the patterns that recur in every ML codebase.The language itself
- Python LibrariesNumPy, pandas, Matplotlib and the rest of the toolkit for data and numerical work.The working toolkit
- SQLOrganise, query and join the data behind an experiment before it reaches a model.Where the data lives
- CUDAParallel computation on the GPU, and the hardware model underneath every training run.Why it is fast
- ROSNodes, topics and services — wiring sensing, communication and control together.Robots, plumbed
PART THREE
Machine learning
Start with what AI is and how to judge a result. Then learn how models use data, how neural networks build representations, and how to investigate their mistakes.
- AI FoundationsWhat the field is, how a system works, and how to judge a result — before any notation.10 chapters
- Machine LearningWeek by week, from cleaning data through models, features and optimisation.5 weeks
- Deep LearningThe longer path through architectures, training dynamics and the things that go wrong.14 weeks
PART FOUR
Simulation & robotics
Make a prediction, change a parameter, and watch what happens. Explore motion, control and sensing in interactive labs, then learn about the environments used to model more complex worlds.