Natural Language Processing

Algorithms and machine learning
Advanced studies
The course will cover state-of-the-art approaches to problems in NLP. We will consider only analyzing written language (not speech). Levels of linguistic analysis we will address: morphology, syntax, semantics, discourse analysis. We will cover standard methods of language analysis that support NLP applications, i.e., rule-based and statistical approaches, and consider some applications in depth: language modeling, bag-of-word models, spell-checking, part-of-speech tagging, parsing. We will use techniques from machine learning: hidden Markov models (HMM) and related algorithms (Viterbi, Forward algorithm, Forward-Backward algorithm), the EM algorithm, unsupervised learning. Students are graded based on 6 compulsory assignments and 2 projects. No exam. Prerequisites: Data Structures and Models of Computation; Strong programming skills; Good knowledge of design and analysis of algorithms; Although the course will introduce the needed mathematical tools, a good level of mathematical maturity is presumed: linear algebra, probability theory, etc.; An understanding of linguistic concepts is required: grammar, word and sentence structure, parts of speech, etc.
Year Semester Date Period Language In charge
2011 autumn 06.09-16.12. 1-1 English Roman Yangarber


Time Room Lecturer Date
Tue 10-12 B119 Roman Yangarber 06.09.2011-12.10.2011
Thu 10-12 B119 Roman Yangarber 06.09.2011-12.10.2011