Ml

Learn Machine Learning: Deep Learning, Seeing and Remembering

Networks for pictures and for sequences, built and measured: convolution and pooling on handwritten digits, recurrent networks and their gates, and models that squeeze and generate.

Start Machine Learning: Deep Learning, Seeing and Remembering free → Browse all subjects

Free to start · adaptive placement finds your level · reviews timed to your own forgetting.

27 lessons in Machine Learning: Deep Learning, Seeing and Remembering

A picture is a grid of numbersMove it one pixelOne small pattern, checked everywhereFilters that find strokesPadding and output sizeMany filters, few weightsKeep the strongest: poolingA small convolution network, trainedPractise on moved digitsClothing is harder than digitsWhat the filters learnedExplanations that explain nothingDeeper, with skip connectionsReuse a trained networkWhen reuse does not payMemory for sequencesRemember the first symbolThe slope that fadesHow far back it remembersGates: choosing what to keepStart ready to rememberSqueeze a digit to two numbersWhat the two numbers keepWhat a t-SNE picture hidesAdding noise, a little at a timeGenerate by removing noiseWhich network for which data

Built to be understood — and remembered.

Every idea is taught with motivation and a worked example before the drills, and an FSRS spaced-repetition engine schedules each review for the day its model predicts you would forget it. A short placement check finds what you already know, so you start Machine Learning: Deep Learning, Seeing and Remembering exactly where it's useful.

Start Machine Learning: Deep Learning, Seeing and Remembering →
Related Ml subjects
Machine Learning: The Whole GameMachine Learning: Learning from Data, in CodeMachine Learning: Neural Networks from ScratchMachine Learning: Build Your Own Language ModelMachine Learning: Doing It Honestly