Topics of Interest Part I

1
Topics of InterestPart I

• Vasileios Hatzivassiloglou
• University of Texas at Dallas

2
Course web page

• http//www.hlt.utdallas.edu/vh/Courses/Fall08/CSP
  roject.html
  
• Up-to-date listing of lectures, schedules, and
  supplemental course material
  
• Also accessible via my home page
  http//www.hlt.utdallas.edu/vh

3
Focus of this lecture

• Discuss several technical areas of interest to
  the instructor
  
• Instructor can describe the state of the art
• Instructor can rapidly evaluate problems and
  proposed solutions
  
• Instructor can suggest modifications to improve
  proposals

4
Focus of this lecture

• Introduce students to several advanced technical
  areas
  
• challenging problems
• cutting-edge technology
• of interest to industry, researchers, and
  government
  
• These are not the only potential areas for
  choosing a project

5
Topics of interest

• Natural Language Processing
• Question Answering
• Intelligent Information Retrieval
• Data Organization and Labelling
• Bioinformatics
• Data and Text Mining
• Medical Informatics

6
Approach

• Heavily empirical relies on massive amounts of
  data rather than human intuition
  
• Based on statistical and machine learning models
  requires appropriate applied mathematics
  knowledge

7
Natural Language Processing

• Natural language processing is the study and
  implementation of computer methods for
  understanding and generating human language
  (English, French, Urdu, etc.)
• The discipline of linguistics studies human
  language mechanisms (as opposed to products) in
  general
  
• Thus, NLP is often referred to as computational
  linguistics

8
Human language

• An extremely complex communication system
• Many symbols (words), many constraints on what
  words can go together, many complications on the
  meaning of combinations of words

9
A comparison with a programming language

• A programming language has a small inventory of
  words with pre-assigned meaning
  
• A programming language has strict rules for
  allowed sequences of words (syntax)
  
• The meaning of each word is unique
• The meaning of a sequence of words is unique
  (compositionality)

10
Levels of NL study

• Word formation from letters (morphology)
• Construction of longer units obeying class
  constraints (syntax)
  
• Interpreting the meaning of words (lexical
  semantics) and longer units (semantics)
  
• Complications from the environment and context
  (pragmatics)
  
• Acoustic properties (phonetics)

11
Teaser

• How does Babelfish translate documents from
  German to English?
  
• How does ETS score SAT essays without the
  computer really understanding what is written?
  
• How does Google automatically populate a calendar
  from email messages?

12
Lexical Semantics

• Lexical semantics is the area of NLP that
  attempts to model, represent, and learn the
  meaning of words
  
• Challenging
• Word meaning is partly compositional
  (derivations, compounds) but mostly arbitrary
  
• Words have multiple meanings
• Correct word usage depends on many hidden factors

13
Components of meaning

• Core meaning
• can be represented using attributes, properties,
  and predicates
  
• Example red
• Variable meaning Frame of reference
• Constraints on word usage depending on context
  (topic, nearby words)

14
Lexical preferences

• Non-substitutable word choices
• red wine
• Restricted patterns of use
• won five straight games
• won five consecutive games
• won five straight
• won five consecutive

15
Semantic polarity

• Deals with a specific aspect of meaning the
  evaluative component assigned to the word
  
• Contrast excellent vs. terrible
• Can be learned automatically by looking at words
  that appear with the word in question

16
Polarity applications

• Separate opinions from factual statements
• Automatically convert text reviews to a numeric
  score (e.g., for movies)
  
• Assess the popularity of a product or person from
  comments (e.g., a cell phone or a politician)
  
• Classify political blogs as Republican or
  Democratic