Cross-Entropy Loss
Domain: Evaluation
🎯 Motivation
The standard loss function for classification. Measures the difference between predicted probabilities and true labels.
📋 Concrete Example
True label: cat (1.0). Predicted: cat (0.7), dog (0.2), bird (0.1). Cross-entropy = -log(0.7) = 0.357. If predicted cat was 0.99, loss = -log(0.99) = 0.01 (much lower).
🔗 Analogy
An archery score - hitting the bullseye (correct class with high confidence) = low loss. Hitting the wrong target = high loss. Cross-entropy measures how far from the bullseye.