The linguistics department of the IEAS has welcomed a new member this term, Niko Schenk. He will present himself in this blog entry.

I am a first-year PhD student in computational linguistics at Goethe University Frankfurt am Main. I have received a B.A. and an M.A. degree in computational linguistics from the Seminar für Sprachwissenschaft at the University of Tübingen.

Currently, I am a member of two separate chairs, led by Gert Webelhuth (IEAS, linguistics) and Christian Chiarcos (ACoLi, computer science).

I am interested in any type of statistical natural language analysis and machine learning techniques, in particular for web data (user-generated content). I’ve spent one year in industry as a software intern at IBM and Sony with a main focus on sentiment analysis.

For a more detailed CV and our open source projects, see my LinkedIn page.

My research is about (implicit) semantic role labeling within the broader context of discourse analysis. In particular, we try to find more abstract (mainly distributional) representations to the underlying semantic structure between any type of discourse units.

In the example,

“John is good in math and sciences. Pail fails almost every class he takes.” (example taken from Marcu Echihabi 2001.)

it is easy to infer that a contrast relationship holds between the two sentences. For a human reader this is indeed “easy”; a computer might need some help… 😉

In this summer term, I teach a Proseminar in grammar engineering and will be teaching another one in practical corpus linguistics, starting next winter term 2015. Students can generally expect hands-on and applied sessions with lot’s of linguistic examples.