Hidden Markov Models (HMMs) for estimating the sequence of hidden states (decoding) via the Viterbi algorithm, and estimating model parameters (learning) via the Baum- Welch algorithm.