Finite-State Methods in NLP

Lecturer(s):Wojciech Skut (Rhetorical Systems Ltd.) and Jakub Piskorski (DFKI Saarbruecken)
Type:Introductory Course
Section:Language and Computation
Time: 17.00-18.30 (Slot 4)
Room:CM G.01


The aim of this course is to introduce the audience to the
finite-state NLP methods. The emphasis is on gaining a good
understanding of several generic finite-state techniques. The
participants learn to apply them to specific problems and to build
their own finite-state applications. The formal background is
introduced gradually and based on concrete examples.

NLP TOPICS COVERED: tokenisation, dictionaries, parsing, text
indexing, tagging. Each topic provides background for several
finite-state concepts and algorithms such as:

	     types of finite-state machines (FSMs): 
	     acceptors/transducers bimachines

	     equivalence operations of FSMs 

	     regular expressions

	     generalised pattern matching

	     rewrite rules
RELEVANCE/TIMELINESS: After decades of intense research, FSMs
constitute one of the most widely used formal frameworks for
NLP. Their applications are very varied, thus the ESSLI audience can
benefit from a uniform introduction covering the typical

PREREQUISITES, LEVEL OF DIFFICULTY: see further particulars.


© ESSLLI 2005 Organising Committee 2005-05-06