www.theBulldogSolution.com

Date and Time: Feb. 12 (Friday), 12-1pm
Place: McF 208
Speaker: Dr. Marjan Mernik

Marjan Mernik received his M.S. and Ph.D. degrees in Computer Science from
the University of Maribor in 1994 and 1998, respectively. His is currently a
professor at the University of Maribor, Faculty of Electrical Engineering and
Computer Science. He is an adjunct professor in the Department of Computer
and Information Sciences at the University of Alabama at Birmingham. His
research interests include principles, paradigms, design and implementation of
programming languages, compilers, formal methods for programming language
description, and evolutionary computation. He is a member of IEEE, ACM and
EAPLS.

Title: MAGIc: A Memetic Algorithm for Grammatical Inference
Abstract: Grammatical inference (grammar induction, language learning), a
subfield of machine learning, attracts researchers in diverse domains such as
pattern recognition, computational linguistics, natural language acquisition, and
software engineering. The goal of grammatical inference is to find a structure
expressed as a grammar from positive (and negative) language examples. The
inferred grammar can then be used for automatic generation of different tools
for language processing (e.g., parser) or be further examined by a software
language engineer with the aim to further enhance the design of the language.
In this talk MAGIc, an evolutionary algorithm that uses local search for
inference of context-free grammars from positive language samples will be
presented.  We have applied MAGIc to infer Domain-Specific Language (DSL)
grammars from DSL programs. Such a scenario is feasible when domain experts
can provide complete DSL programs or excerpts of such programs. I will show
several examples of DSLs (e.g., DESK, HYPERTree) for which grammars have
been successfully inferred using MAGIc.