Linguistics 581: Computational Linguistics
Concept map of the Course
Words & Finite-state methods
String matching and search
- Relation of
regular languages to finite-state automata (FSAs)
- Finite-state transducers (FSTs): A special
kind of FSA for finite state relations
- Morphology: Breaking words
into their pieces
- Morphology and FSTs
- Spelling
Correction and Minimun Edit Distance Computing (Viterbi I)
Ngram models and word occurrence models [and Data!]
- Introduction
to probability, conditional probability and entropy
- Bigram and trigram models: Bigram models
are Markov chains
- Smoothing probabilistic
models
- Corpora: Online
collections of annotated data
- Part of speech tagging
- Hidden Markov Models
(HMMs) and part of speech tagging
HMMs, Max Ent Models, and MEMMs
- Introduction
to HMMs
- HMMs
in Part of Speech Tagging
- Decoding HMMs
(Viterbi II)
- Linear regression,
logistic regression
- Maximum entropy
models
- Maximum entropy
models and sequential probability (Max Ent Markov Models, MEMMs)
Syntactic models and parsing
- Context free grammars
- Inadequacy of FSAs
for context-free languages
- Application of context free
grammar to language description
- Parsing context-free grammars
- Probabilistic models
of syntax: Probabilistic context-free grammars
Computational semantics: Word Meaning
- Word senses
- Vector space model
of words
- Meaning similarity
measures