Integrals appear in expected value computation, marginalizing over latent variables, AUC-ROC calculation, and Bayesian inference where posterior distributions require normalizing constants.

Expected value of a continuous random variable: E[X] = ∫ x dot f(x) dx. AUC (area under the ROC curve) is literally an integral - the probability that a random positive sample ranks higher than a random negative. Bayesian marginal likelihood: P(data) = ∫ P(data|theta) P(theta) dtheta - usually intractable, motivating variational inference and MCMC approximations. In practice, ML practitioners encounter integrals mainly through probability theory rather than direct computation.