433-460 Human Language Technology

Credit may not be gained for both 433-460 Human Language Technology and 433-660 Human Language Technology.

Study at the third-year level in at least four of the following areas: artificial intelligence, computer design, database systems, graphics, interactive system design, networks and communications, operating systems, programming languages and software engineering, and theory of computation. Completion of 620-201 Probability (or equivalent) would be an advantage.

The objectives of this subject are for students be familiar with the foundations of symbolic and statistical natural language processing; be familiar with key concepts in language desription and analysis; be able to develop and evaluate computational models of language; and be familiar with a variety of human language technologies.

Topics covered include the linguistics of words and phrases, part-of-speech tagging, finite-state transducers, chart parsing and chunk parsing, hidden Markov models, n-gram language models, spelling and grammar checking, collocation analysis, word-sense disambiguation, text retrieval, information extraction, and machine translation. Programming work will be undertaken in the Python language, and will use NLTK, the Natural Language Toolkit (nltk.sf.net).

Four projects, expected to take about 36 hours, during semester (50%) and a 2-hour end-of-semester written examination (50%). To pass the subject, students must obtain at least 50% overall, 25/50 in assignments and project combined, and 25/50 in the written examination.