Concept Map: Machine Learning

Each Machine Learning lesson teaches a handful of ideas; the project and the final assessment ask you to use them together. This map shows how they connect. Each concept sits in the row of the theme of the lesson where it first appears (F, D, U, R, C, M, E, in course order) and in the column of the workflow stage it belongs to: the problem and delivery, data and splitting, the model, evaluation. The colors are the same as in the workbook, the cheat sheet and the study cards.

Four ways to use it:

• Tap a concept. Its neighbors light up and the panel shows what it is, the lessons where it appears (linked to the section) and, one sentence each, where it leads and where it comes from. Tap a neighbor to follow the thread.
• Follow a trail. Each trail walks through one of the course's questions, step by step, such as "why can 95% accuracy be worthless?", "from text to vector" or "from the notebook to production".
• Pick a lesson. The buttons are grouped by theme; they light up the concepts the lesson introduces and the ones it revisits from earlier lessons. The U4 and U5 buttons include their sub-pages.
• Search. The search box finds a concept by name, in English or Portuguese, or by its definition (Enter opens the first match).

On a phone, the concepts are grouped by theme, the lessons sit under one tab per theme and the panel slides up from the bottom of the screen; close it with ×.

To see everything at once, on paper, use the cheat sheet (in Portuguese). To practice, the study cards and the workbook (in Portuguese).