HPSG-based Syntactic Treebank of
Bulgarian (BulTreeBank)
The main objective of BulTreeBank Project is
to create a high quality set of syntactic structures of Bulgarian sentences
within the framework of HPSG. Ideally, the tree bank should contain samples of
all the syntactic structures of the language. These sentences should serve as
templates for future corpora development, could become the basis for the
development of a more comprehensive test suite for NLP applications, can be used
as a source for grammar extraction and for linguistic research. The development
of such a tree bank is inextricably linked to the development of a formal
grammar and a parser for Bulgarian. We propose to write two grammars for
Bulgarian, none of them exhaustive, but in our view a necessary stage on the way
to a complete formal grammar of the language: one imposing very general
constraints on the syntactic structure and one for partial parsing. Within the
project we will prolong the development of CLaRK system - XML-based
system for corpora development.
The core of CLaRK is an XML Editor, which is
the main interface to the system. Besides the XML language itself, we
implemented an XPath language for navigation in documents and an XSLT language
for transformation of XML documents. CLaRK is based on an Unicode encoding of
the information inside the system. The basic mechanism of CLaRK for linguistic
processing of text corpora is the cascaded regular grammar processor. Several
mechanisms for imposing constraints over XML documents are available. The
constraints cannot be stated by the standard XML technology.
Research associate in Computer Science and Computational Linguistics
Dimitar Doikoff
Master's student in Computational Linguistics
Ilko Grigorov
Master's student in Computer Science
Hristo Ganev
Master's student in Computer Science
We would like to thank Gergana Popova (PhD student, Department of Language
and Linguistics, University of Essex) and Atanas Kiryakov
(Head of OntoText Lab.) for their help in
preparation of the proposal of the project.
We would like to thank Dr. Frank Richter from SfS for his contribution to our better
understanding of the last trends in HPSG during his visit at the Linguistic Modelling
Laboratory at the beginning of the BulTreeBank Project.
Expected Results
A set of Bulgarian sentences marked-up with detailed syntactic
information. These sentences will be mainly extracted from authentic Bulgarian texts. They
will be chosen with regards two criteria. First, they will have to cover the variety of
syntactic structures of Bulgarian. Second, they we show the statistical distribution of
these phenomena in real texts.
Inside the TreeBank a core set of sentences will be designated. The purpose of this core
set of sentences is to serve as a test-suite for software application processing texts on
Bulgarian on the level of syntactic descriptions.
Reliable partial grammar for automatic parsing of phrases in Bulgarian. This grammar
will be extensively tested and used during the creation of the TreeBank. It will be used
as a module separate from the TreeBank in tasks which require only partial parsing of
natural language texts such as Information Retrieval, Information Extraction, Data Mining
from Texts and etc.
Software modules for compiling, manipulating and exploring the data in the TreeBank.
This software will support both the creation of the TreeBank, and its use for different
purposes such as automatic extraction of grammars for Bulgarian.
Contacts
Kiril Simov
BulTreeBank Project
Linguistic Modelling Laboratory, CLPP,
Bulgarian Academy of Sciences
Acad. G.Bonchev St. 25A
1113 Sofia, Bulgaria
Tel: (+359 2) 979 2825
Fax: (+359 2) 70 72 73 www.BulTreeBank.org