Computational Lexicography and Grammar

Summary

Natural language processing (NLP) of Slavic languages; Lexical databases; Sorting; Exporting and importing; Morphological parsers; Morphological generators.

Homework

HW # 8 (due end of Week 12): Create a 50- entry bilingual dictionary (non-Western language X - English) with equivalents, POS tags, usage labels, examples and their translation in the form of a Microsoft Access database. E-mail both the database and the English – non-Western language X index as a text file exported from the database.

NLP of Slavic Languages

1933 USSR patent by Troyanski
GAT 1952
ALPAC Report 1966
SYSTRAN
CETA, SUSY, ETAP, ViRTEX, METAL
APAC, RUSLAN
NeuroTran

Electronic vs. Traditional Lexicography and Grammar

Preciseness
Comprehensiveness
Testability
Role of the corpus
Queries

Dictionary Microstructure

Entry word
Lemma
Definitions/equivalents
Examples

Morphosyntactic rules

Analyzers vs. generators
Overgeneration, undergeneration
Parsing rules
Transfer rules

The following three about the machine translation system NeuroTran will give you some ideas about theoretical background, morphological generator, and syntactic component of an MT system.

Lexical Database - An Example

Wordnet, EuroWordNet, Global Wordnet, BalkanNet

Working with databases

Exporting, importing
Sorting
Filtering
Creating masks

Perl comes next - remind yourself of Java Script

Before we start working on serious stuff in Perl, it is useful to remind yourself of some functions in Java Script. Here are three flashy pages:

Look, I can detach my head and have it levitate
This page flies
Rotating banners luring inocent students into my courses

If you like folklore, this is cool JavaScript Story generator based on Propp's theory.

Example

Last year's example