03·Three learning paradigms: supervised, unsupervised, reinforcement

Introduction

When a developer says "I am training a model", in 95% of cases they are doing it in one of three paradigms. Supervised learning — learns from (input, correct answer) pairs, like a student with an answer key. Unsupervised learning — looks for structure in data without hints, like an anthropologist describing an unknown tribe. Reinforcement learning — learns by trial and error, receiving rewards and penalties, like a child learning to ride a bike. In this lesson you will learn to distinguish these three approaches from real-life examples and to pick the right one for the problem.