TOYOSHIMA Dojo
@
Theoretical linguistics is my specialty and I am investigating the computational system of human language, how language is generated and processed in the human brain, based on mathematical modeling. I am particularly focusing on the economy issues in theoretical linguistics, that is, how to reduce the computational complexity implicated in the theory of generative grammar, and its viable implementation in the human brain. In the interdisciplinary field of computational linguistics in recent years, applied linguistic engineering researches are vigorously carried out to develop speech dialogue systems and machine translation systems. In my laboratory, we accommodate research themes to fit the students' needs and interests as much as we can. For example, one student is currently exploring whether and how much it is possible for artificial neural networks to learn grammar through sequences of binary numbers converted from natural language, in a setting just as human infants acquire language naturally without explicit instructions.

I frequently come across engineering students who naively assume that it is a matter of computational resource to design a speech interface for a machine, say a robot, without being aware of the complexity and problems inherent in natural languages. Therefore, I have prepared this crash course to impart the basic understanding of language, necessary for applied researches in linguistic engineering, including voice recognition, voice synthesis, machine translation, data mining and dialogue systems. In the first half, students are to learn the basics of linguistics from phonetics to semantics, familiarizing themselves with the terminology and concepts for the literacy in the field. In the latter half, I will provide personalized instructions tailored to each student's chosen research topic. Through this package, I hope that students would be able to develop a concrete research plan, identifying what are the problems, where a solution can be found, and how they can solved, in order to reach their chosen research objectives.



To Page Top

@