MIME-Version: 1.0 Server: CERN/3.0 Date: Monday, 16-Dec-96 23:12:34 GMT Content-Type: text/html Content-Length: 9413 Last-Modified: Friday, 12-Apr-96 22:27:57 GMT CS674 Spring 1996 Course Materials

CS674 Spring 1996
Introduction to Natural Language Understanding
Course Materials



Handouts

(Tentative) Course Syllabus
Course Description and Policies

Lecture Slides

Introduction to Natural Language Understanding

(Jan 22) Introduction to the Field of NLP (Ch 1)
(Jan 24) Stages of Processing

Syntactic Analysis

(Jan 29) Grammars and Sentence Structure, Top-down and Bottom-up Parsing (Ch 3.1-3.3)
(Jan 31) Chart Parsing (Ch 3.4)
(Feb 5) Feature Systems and Augmented Grammars (Ch 4.1-4.5)
(Feb 7) Human Preferences in Parsing, Partial Parsing (Ch 6.1, 6.5)

Semantic Analysis

(Feb 12) Word Senses and Ambiguity,Representing Verbs and States (Ch 8.1-8.6)
(Feb 14) Thematic Roles, Semantic Interpretation (Ch 9.1-9.4)
(Feb 19) Selectional Restrictions, Handling Intrasentential Word Sense Ambiguity (Ch 10.1--10.2)

Conceptual Sentence Analysis

(Feb 21) The CIRCUS Parser, Preference Semantics, Data-Driven Semantics

Context and World Knowledge

(Feb 26) The Problem of Inference, Expectation-Based Text Analysis (Ch 15.1-15.7)
(Feb 28) Using Knowledge About Action and Causality, Scripts
(Mar 4) Plan-Based Understanding of Text
(Mar 6) Discourse Context, History Lists, Centering (Ch 14.1-14.3) ***Guest lecture***
(Scott Mardis)

Current Trends: Evaluation, Learning, Statistics

(Mar 11) Evaluating NLU Systems
(Mar 13) HMM's (Ch 7.1-7.4)
***** Spring Break *****
(Mar 25) Part-of-Speech Tagging
(Mar 27) Probabilistic Context-Free Grammars, Best-First Parsing (Ch 7.5-7.7)
(Apr 1) Context-Dependent Best-First Parsing, Statistical Word Sense Disambiguation (Ch 10.4-10.6)
(Apr 3) A Localist Connectionist Approach to Sentence Analysis
(Apr 8) Transformation-Based Error-Driven Learning and NLP
(Apr 10) Corpus- and MRD-based Methods for Word-Sense Disambiguation
***Guest lecture*** (Julia Komissarchik)
(Apr 15) Information Extraction as a Basis for High-Precision Text Categorization
No on-line slides for this lecture.
(Apr 17) A Case-Based Approach to Ambiguity Resolution
(Apr 22) NO CLASS
(Apr 24) Project Presentations
(Apr 29) Project Presentations
(May 1) Project Presentations

Homework Assignments


Project Information

What to Turn in for the Proposal


What to Turn in for the Project

Official due date: Friday 5/3. We'll accept projects without penalty until 5:00, Friday 5/10.

Programming projects:

Final Writeup for Programming Projects (just a few pages):
  1. Problem description
  2. Description of general approach
  3. Description and results of evaluation
  4. Discussion. (what worked; what didn't work; options you'd like to have tried; analysis of the results; etc.)
  5. I'd also like to see any code that you wrote and a few short traces of the system in action (if that makes sense for your project).

In-Class Presentation: ~7 minutes in length. Should include an overview of the problem; your solution; evaluation method and results.

Non-Programming projects:

Final Writeup for Non-programming Projects. This will vary for each person, but in general, the final writeup for non-programming projects will probably contain a description of the problem that you looked at; a summary of the papers that you read; a critique of the existing approaches; your attempt at an evaluation of the the theory/algorithms presented in the papers on real text.

In-Class Presentation: For non-programming projects, the in-class presentation should be a synopsis of what you'll include in the paper. You'll have to leave a lot out of course...


Programming Projects from Last Year's Class

Reading Projects from Last Year's Class


Return to CS674 home page