MIME-Version: 1.0 Server: CERN/3.0 Date: Monday, 16-Dec-96 23:11:20 GMT Content-Type: text/html Content-Length: 5471 Last-Modified: Friday, 12-Apr-96 14:11:28 GMT CS674 Project Presentation Schedule

CS674 Project Presentation Schedule

April 24th

  1. Kendra Willson
    Thematic roles and NLU
    This paper muses on issues surrounding the adaptation of various linguistic approaches to argument structure for NLP purposes, with the ultimate but oblique goal of shedding light on the age-old question: Are linguists useless?

  2. Catherine Starkey
    A German/English Translator for Simple Sentences
    This project will take as input a series of sentences in English or German and then parse them according to a corresponding set of grammar rules. Then, using a different set of rules (for the other language), it pieces together the new sentence with words entered in a lexicon.

  3. Lin Hsian Wang
    Prepositional Phrase Attachement by Rule-Based Approach
    To solve the prepositional phrase attachment ambiguation using set of rules generated by Transformation-Based Error-Driven Learning algorithm.

  4. Mao Yonghong
    Part of Speech Tagging Algorithms
    I will compare some part of speech tagging algorithms and try to give some suggestions.

  5. Eric Scharff
    Neural Networks for Part of Speech Disambiguation

April 29th

  1. Alfred Hong
    Translation of Regular Query Sentences into SQL Queries
    The idea of this stems from the numerous Web forms today that provide search capability. Often times, one would like to have a query interface to a database (relational), but the search definition parameters are usually restrictive and "inhuman." Adding a natural language query capability to the search function while providing a back-end conversion to an equivalent SQL query to a database would makes things more human and natural and hopefully more powerful.

  2. Ed Wayt

  3. Jonathan Decristofaro
    Making the Chart Parser Avoid Useless Work
    The bottom-up chart parser can be very inefficient when there are many ways to parse a phrase. This work eliminates the multiple parses generated when there is an ambiguous phrase (like a compound noun of five words).

  4. David Walker
    Statistical Methods for Part-of-Speech Tagging

  5. Wee-Liang Heng
    Probabilistic Part of Speech Tagging
    This project explores various statistical models (e.g., smoothed bigram language model) for part-of-speech tagging, and implements the associated algorithms. The models will be evaluated on their accuracy on either the PennTree Bank or Brown corpus.

  6. Daniel Brown
    Text Compression Using NLP
    Most compression schemes take advantage of the fact that certain symbols in the input file are much more likely (frequent) than others. This project will attempt to take advantage of the fact that certain sequences of words are much more likely than others. In particular, I will assume that grammatical sentences are much more likely than ungrammatical sentences. I will encode sentences in such a way that grammatical sentences take up less space.


May 1st

  1. Heji Kim
    Information Extraction from Poetic Forms
    The purpose of this project will be to automatically gather useful information from poems via syntactic/semantic parsing; this information will be the search keys to index a corpus of poetry in a database.

  2. Vera Kettnaker
    Deriving the Meaning of Novel Words
    The idea of this project is to guess the meaning of novel words (first names, abbreviations of new terrorist organizations ect.) on the basis of expectations we gain from the surrounding context.

  3. Grzegorz Czajkowski

  4. Pavel Naumov
    Presentation of Natural Language Sentence in Nuprl Term Editor
    Adopt Nuprl term editor for natural language representation. Write several ML programs that will extract information from natural sentences presented in term editor. (say, answer on simplest questions base on given group of sentences).

  5. Joel Rosenzweig
    A Weather Text Processing and Understanding System
    I am writing a grammar suitable for parsing weather text that I can grab off of the Internet/Weather Channel Web site. This system will look for "thematic" roles associated with the text to provide a summary of the forecast. Time permitting, I will make a program that will interpret this summary data, and generate a weather map that depicts the forecast.

  6. Christine R. Paradis
    Translating an English Sentence to an SQL Statement
    I will be translating an English sentence to an SQL query. My initial approach to this problem is to use the thematic role information for a given sentence and map this information to the corresponding SQL constructs. For evaluation of the system, I will generate SQL statements for a set of test sentences and compare the results with my desired/correct SQL statements.