COSC 6397 Natural Language Processing

Fall 2014

MW 1:00 - 2:30 PM, SW 221

Instructor: Arjun Mukherjee


Overview

This is a introductory natural language processing course (NLP). The course in intended for developing foundations in NLP and text mining. The broader goal is to understand how NLP tasks are carried out in the real world (e.g., Web) and how to build tools for solving practical text mining problems. Throughout the course, emphasis will be placed on a understand NLP concepts and tying NLP techniques to specific real-world applications through hands-on experience. The course covers fundamental topics in statistical machine learning and touches upon topics in sentiment analysis and psycholinguistics.


Administrative details

Flyer
Syllabus
Instructor office hours: MW 2:30 - 3:30 PM, PGH 582

Prerequisites

The course requires basic background in mathematics and sufficient programming skills. If you have taken and did well in one or more of the equivalent courses/topics such as Algorithms, Artificial Intelligence, Numerical methods, or have some background in probability/statistics, it will be helpful. The course however reviews and covers required mathematical and statistical foundations. Sufficient experience for building projects in a high level programming language (e.g., Java) will prove beneficial.

Note: This course has minor overlap with Data Mining (COSC 6335) and Machine Learning(COSC 6342). Especially some topics in supervised learning as they lay the foundation for other NLP algorithms to be covered in this course. Hence, they are covered to make this course standalone. Although not required, however, if you have taken either of those courses, it will be helpful.

Reading Materials

Textbooks:

SI: Statistical Inference, Casella and Berger. Cengage Learning; 2nd edition.
FSNLP: Foundations of Statistical Natural Language Processing, Chris Manning and Hinrich Schütze. MIT Press. Cambridge, MA: May 1999. Companion website for book.
WDM: Web Data Mining: Exploring Hyperlinks, Contents, and Usage Data. Bing Liu; Springer, 1st Edition.

Required reference materials:

Online resources (OR) per topic as appearing in the schedule below.
Lecture notes

Course materials (slides, lecture notes, etc.) (You may use 7zip to unpack)

NLP tools (POS Tagger, Chunker, Naive Bayes, etc.) and templates with linked libraries for research project

Assignment Due Dates and Grading

Component Contribution Due date
HW1 10% 09/03
HW2 10% 09/17
HW3 10% 10/06
HW4 10% 10/22
HW5 10% 11/10
Mini Project 25% 09/29
Optional Research Project 50% 12/08
Final 20% 12/05, 1-4 PM, SW 221
Class contribution 5% -


Rules and policies

Late Assignments: Late assignments will not, in general, be accepted. They will never be accepted if the student has not made special arrangements with me at least one day before the assignment is due. It also needs to be a justifiable reason owing to exacting circumstances. If a late assignment is accepted it is subject to a reduction in score as a late penalty.
Cheating: All submitted work (code, homeworks, exams, etc.) must be your own. If evidence of code sharing is found, you will receive an F grade in the course. Please refer to the student handbook for details on academic honesty.
Statute of limitations: Grading questions or complaints, will in general not be attended to beyond one week after the item in question has been returned.


Schedule of topics

Please note that the following is a list of tentative topics. During the course, and interleaved between lectures, time will be invested in review questions, homework solutions, discussion of novel ideas, project updates, and concept review before exams.

Topic(s) Resources: Readings, Slides, Lecture notes, Papers, Pointers to useful materials, etc.
Introduction
Course administrivia, plan, goals, NLP Resources
Language as a probabilistic phenomenon, Zipf's law
Word collocations and text retrival
Required readings:
Lecture notes/slides
Chapter 1 FSNLP (Sections 1.2.3, 1.4, 1.4.1, 1.4.2, 1.4.3, 1.4.4)
Boolean retrieval slides by H.Schutze
Boolean retrieval [Manning et al., 2008] (upto section 1.4)
Statistical foundations I:Basics
Probability theory
Conditional probability and independence
Required Readings:
Lecture notes/slides
Chapter 2 FSNLP (Section 2.1.1 - 2.1.10)
Chapter 1 SI (Full reading recommended. Focus on topics covered in class and solved examples)
OR01: X.Zhu's notes on mathematical background for NLP
Statistical foundations II: Random varibales and Distributions
Random variables, density and mass fuctions
Mean, Variance
Common families of distributions
Multiple random variables: joints and marginals
Required Readings:
Lecture notes/slides
Chapter 2 SI (Theorem 2.1.10, 2.2, 2.2.1, 2.2.2, 2.2.3, 2.2.5, 2.3.1, 2.3.2, 2.3.4, and topics covered in class).
Chapter 3 SI (All sections + worked out examples upto 3.4), only distributions that were covered in class.
Chapter 4 SI (4.1, 4.1.1, 4.1.2, 4.1.3, 4.1.4, 4.1.5, 4.1.6, 4.1.10, 4.1.11, 4.1.12, 4.2.1, 4.2.2, 4.2.3, 4.2.4, 4.2.5).
OR02: K.Zhang's notes on common families of distribution with worked out examples [Skip hyupergeometric, neg-binomial, geometric distributions and read only those covered in class].

Optional Recommended reading/solved examples:
OR03: Notes on Joint, marginals, worked out examples by S.Fan
OR04: Tutorial on joints and marginals by M.Osborne [Contains NLP specific examples]
Words
Collocations
Hypothesis testing, statistical tests, p-values
N-gram language models
Required Reading(s):
Chapter 5 FSNLP (5, 5.1, 5.3, 5.3.1, 5.3.3), Chapter 6 FSNLP (upto 6.2.2).
OR07: J.Zhu's notes on t-test
OR08: Lecture notes/slides on N-gram langauge models by Y. Choi: (1), (2)
OR09: Recommended readings: (0) Chapter 6 of [Jurafsky and Martin] (1) Langugae modeling notes by M.Collins (2) Lecture notes by K.Mckeown
t-table, Chi-square table
Markov models and POS tagging
Hidden markov model (HMM)
Part of speech tagging
Required Readings:
Chapter 9 FSNLP (upto 9.4), Chapter 10 FSNLP (upto 10.2.2)
OR10: (1) Lecture notes/slides by M. Marszalek on "A Tutorial on Hidden Markov Models by Lawrence R. Rabiner", (2) Toy problems by E.Lussier, (3) POS Tagging by Y.Choi

Programming resources, tools, libraries for projects and homeworks:
(1) HMMs and sequence taggers
JAHMM: Implementation of an HMM in Java, Mallet, SVMHMM, CRF++
(2) POS Taggers
OpenNLP, Stanford Parser, (Online version), Illinois Chunker, POS Tagger
Grammar and Parsing
Shallow Parsing and Phrase Chunking
Context Free Grammars (CFGs)
Top-down and Bottom-up parsing
Probabilistic Context Free Grammars (PCFGs)
Statistical Parsing and PCFGs
Required Readings:
Lecture notes/slides
Refer to Chapter 9, 10 and 12 of this book for select topics covered in class

Programming resources, tools, libraries for projects and homeworks:
Chunkers, shallow and full parsers: (1)OpenNLP, (2) Stanford Parser, (Online version), (3) Illinois Chunker, POS Tagger
Text categorization
Decision trees
Naive Bayes
Support Vector Machines
Evaluation metrics
Significance testing (revisited)
Feature selection schemes

Required Readings:
Chapter 3 WDM (3.1, 3.2, 3.6, 3.8, 3.3)
F. Keller's tutorial on Naiye Bayes + notes of A.Moore for graph view (Slide 8)

Programming resources, tools, libraries for projects and homeworks:
SVMLight
LIBLINEAR (Implements L1/L2 regularized classification and regression using SVM, SVR, LR with support for L1/L2 loss), Mallet, LingPipe, C. Borgelt's DM Tools
Sentiment Analysis and Psycholinguistics
Aspect extraction
Deception and opinion spam
Required Readings:
Lecture notes + slides + selected topics (covered in lectures) from Chapter 11, WDM
Paper on opinon spam: [Ott et al., 2011], slides, demo

Programming resources, tools, libraries for projects and homeworks:
Pos/Neg Sentiment Lexicon, SentiWordNet, Deep learning for senitment analysis