| Language
Technology
Led by: Timo Honkela (Aalto University) and Krister
Lindén (University of Helsinki)
The Language Technology sub-programme is multidisciplinary, as
a reflection of the subject itself. At the core of language technology
is the computer modelling of natural language, giving rise to a
variety of applied research topics. These include the development
of natural language parsing methods for the recognition of natural
languages, as well as the generation of natural languages (in principle,
natural languages cover all human languages, whether written, spoken
or signed).
Language technology (previously referred to as computational linguistics)
involves the modelling of language recognition and generation, as
well as modelling language structure, function and development,
for example through syntactic analysis, computational models of
grammar, computer-based lexicons and the methods for their generation.
Applied research topics include:
- tools for writers, such as spell check, hyphenation and grammar
check
- information retrieval, information extraction, document indexing
and term recognition within running text
- machine translation and other translation tools
- interfaces in natural languages where a person communicates
with a computer (either by speaking or writing) and the related
dialogue studies
- computer-assisted language learning, where the structure/grammar
of the target language has been modelled on the computer or where
the learning materials are derived and refined by language technological
means.
- speech synthesis and speech recognition methods that take account
of the essentials of a language
Not all language-related computational applications are actually
language technological. In word processing, for example, the writer
writes in a language –which the programs process and save–without
the word-processing program being considered a language technological
application. Only functions such as a grammar check, which models
grammar and dictionary information, is language technology proper.
A language-learning program could equally well be like an audio
and/or a video recorder in its essentials, with no dependence on
any functioning information as such or modelling of linguistic structures.
The point at which the application becomes language technological
is when information on the functions and structure of the language
is incorporated in the program and the program can guide the learner
in the right direction.
|
