Abstract:
The initiation of this thesis is how to understand the natural language that contains an enor-
mous number of ambiguity aspects. With the aim of observing difficulties and coping with
them, a system will be developed by applying natural language understanding (NLU). How-
ever, dealing with the natural language demands a lot of processing, word, syntax, semantic
analyzing and etc. Most of these tasks are successful and can be used in many applications, a
search engine for example. The most difficult task of Natural Language Processing is Natural
Language Understanding since the correct meaning has to be precisely determined and there
are a huge number of possible meanings. Therefore, the focus of this thesis is this obscure
component. The system going to be developed is a pipeline with several subsystems taken
from other publicly available sources except some of the NLU part. The understanding part
will be deliberately examined in order to disambiguate any ambiguities and yield the result
the user needs. To make it feasible to solve, NLU will be applied to a small specific domain,
the academic interview domain in which the vocabulary and number of word meaning size
shrink and the language structure is much simpler. A number of sample AIT applications
will be used to find out the language specification. Eventually, essential knowledge gained from this thesis will potentially be useful in a bigger domain.
