Natural Language Processing

What you’ll study

Words, parts of speech, syntax, grammars and semantics; language modelling in general and the noisy-channel model. Linguistics — phonology and morphology, word classes and lexicography. Mutual information, the t-score and the chi-square test. Hidden Markov models — the trellis and the Viterbi algorithm; HMM tagging (supervised and unsupervised) and evaluation methodology — precision, recall and accuracy. Statistical, transformation-rule-based, maximum-entropy and feature-based tagging, with results on tagging various natural languages. Non-statistical parsing algorithms — a simple top-down parser with backtracking — and probabilistic parsing. An introduction to statistical machine translation.