CPSC 503 Computational Linguistics

1
CPSC 503Computational Linguistics

• Course Check Point
• Lecture 7
• Giuseppe Carenini

2
Today 4/2

• Introductions
• Course Objectives
• Projects Overview
• Working with Corpora
• Pearl (motivation and pointers)
• Summary so far

3
Introductions

• Your Name
• Previous experience in NLP?
• Why am I interested in NLP?
• Are you thinking of NLP as your main research
  area? If not, what else do you want to specialize
  in.
• Anything else

4
Course Objectives

• Become familiar with current research in NLP to
  be able to read most recent papers and get a
  high-level
• Understanding of key problems and solutions
• Understanding of methodology
• In your project area, to be able to read recent
  papers and understand the details
• Devise and implement (small) extensions in
  meaningful directions

5
Final Project

• Critical review and extension of recent research
• Read several papers about it
• Either improve on the proposed solution (e.g.,
  using more effective technique)
• Or propose new solution
• Write report discussing results
• Present results to class
• These can be done in groups (max 2).
• Read ahead in the book to get a feel for various
  areas of NLP

6
Final Project Topics (1)

• Query expansion for intelligent web search
• e.g.,I am looking for a car mechanic in
  Yaletown
• (car mechanic OR garage) AND ((downtown
  Vancouver) OR Yaletown))
• Summarization (e.g., summarize a series of news
  articles of the same event)

7
Final Project Topics (2)

• Generation of evaluative arguments (e.g.,
  comparing several entities / tailored to several
  people advising 2 people on which two countries
  to visits)
• Dialog Systems (preferably Human-Robot
  Interaction) (e.g., Robot assisting users with
  physical impairments to fetch and deliver objects
  in office environment)

8
Final Project Topics (3)

• Analyzing linguistic expression of attitudes
  (opinions and evaluations) in text - e.g.,
  Recognizing and Organizing opinions in World
  Press
• Text reconstruction - e.g., reconstructing an
  email from partially ordered quotations of it)
• Analysis of Collocations (Chp. 5 MS - see
  readings) stiff breeze ?stiff wind?
• Limited compositionality, hypothesis testing
• Used in IR, NLG end Lexicography.

9
Final Project Topics (4)

• Generating arguments to change people behavior -
  e.g., Convince people to quit smoking
• Other options
• Projects on languages other than English (
  caveat you might not find corpora and code)
• For other options check list of papers at course
  Webpage (I got it from prof. at UPitt)
• Key requirement you have to think hard about how
  you will evaluate your work

10
Final Project Tasks

• Narrow down possible topics ASAP
• Then talk to me about it
• Mar 3rd Project Proposal (1 page)
• By end of course Apr 10-20
• Write Report
• Present results to class (15-20min presentation)

11
Corpus-Based Work

• Copy of Chp. 4 MS in reading room
• Raw Data (plain text)
• What is a word? Tokenization
• What is a sentence?
• Marked up data (by hand, automatically, or
  mixture)

12
Perl Motivation

• (Used) to be the preferred language for text
  processing
• Powerful embedded RegExp
• Optimized for string manipulation
• Easy file handling
• Lots of Perl code out there for text processing
  and NLP
• Flexibility(), Portability, Extensibility and
  Free!

13
Learning Perl Objectives

1. To be able to write relatively simple Perl
   programs (e.g., for your project you may get a
   corpus that needs some heavy restructuring)
2. To be able to understand and modify relatively
   complex Perl programs (e.g., for your project you
   may find code that does something similar to what
   you need but not exactly)

14
Perl Key Ideas

• Easy to learn how to do simple things
• Very flexible (use comments!)
• Meaning of expression depends on context
• Many, many built-in functions (Ill send you a
  pointer)

15
Perl Readings

• Required
• Read online cs219 Perl CourseNotes (max 3h)
• Suggested (for more examples)
• Read online Picking Up Perl (max 2h)
• Read online D. Inkpen Perl Tutorial (max 1h)
• Read excerpts from Programming Perl (copy in
  reading room). (max 2h)

16
Perl To Do List

• Download and study simple programs from course
  Web page.
• Work on assignment-2
• Perl warm-ups,
• Perl text processing,
• Perl info extraction,
• Perl min-edit-distance.

17
Course checkpoint

• NLP
• The study of formalisms, models and algorithms to
  allow computers to perform useful tasks involving
  knowledge about human languages.

18
Spelling multiple errors

• (1) Given O, for each wi compute
• meimin-edit distance(wi,O)
• if meiltk save corresponding edit operations in
  EdOpi

(3) Sort and display top-n to user
19
Next Time

• Chp.6 N-Grams (assign probabilities to sequences
  of words)
• Assignments