Backpropagation
Domain: AI Fundamentals
🎯 Motivation
Calculates how much each weight contributed to the error, so we can adjust them. The learning step.
📋 Concrete Example
The model predicted cat with 60% but the true label is dog. Backprop computes gradients telling each weight how to change.
🔗 Analogy
Like a coach reviewing a lost game with each player, telling them exactly what to fix for next time.