Home Topics About

Theme

Loss Functions

Understand how probability distributions become model penalties.

  1. 01 From Probability Distributions to Loss Functions Learn how maximum likelihood connects distributions to loss functions, why raw likelihood underflows, and how negative log-likelihood makes training possible.
  2. 02 Normal Distribution to Mean Squared Error Loss Learn why Normal distribution errors lead to mean squared error loss, using a basketball story to explain MSE, squared errors, and big-miss penalties.
  3. 03 Laplace Distribution to Mean Absolute Error Loss Why Laplace distribution errors lead to mean absolute error loss, using mansion-priced outliers to explain MAE, fat tails, and calmer penalties.
  4. 04 Bernoulli Distribution to Binary Cross-Entropy Loss Why Bernoulli yes/no outcomes lead to binary cross-entropy loss, using rain probabilities to explain two buckets, confident wrong guesses, and negative log penalties.
  5. 05 Categorical Distribution to Categorical Cross-Entropy Loss Many labels. One winner. Learn why categorical cross-entropy is the natural loss for many-choice classification and how the correct bucket decides the penalty.
  6. 06 Poisson Distribution to Poisson Loss Learn why Poisson loss is the surprise meter for counts, rates, and fixed windows—from failed logins to any event that arrives one tally at a time.