Homepage of Björn Rudzewitz

I'm working in the project A4 in the SFB 833 as a researcher on the question of how content assessment for learner data can be performed by computers. This question is especially interesting since language learners, in contrast to native speakers, have only acquired an intermediate state of proficiency in their L2. This results in a high degree of variability in terms of form and meaning in utterances produced by language learners. In the A4 project we explore computational approaches for dealing with this variation in order to assess the semantic content of language learner data.

Besides working on the technical aspects of the CoMiC system (e.g. porting it to uimaFit and DKPro, ...), I wrote my B.A. thesis (link) about how general linguistic properties and the explicit task context can be used to improve short answer assessment of learner language in a reading comprehension scenario there. In addition I successfully demonstrated the ability of the system to adapt to new domains such as Question Answering, Answer Selection, and Plagiarism Detection. This and further work in the context of this project is documented in several scientific articles.

Furhtermore I'm working as lead developer on an interactive school workbook with feedback generation (FeedBook) in SFB 833 Transfer project 1 (T1) to refine my technical abilities and gain experience in coordinating bigger software projects. In this project we explore how online scaffolding feedback can be generated in an intelligent, interactive workbook for learning English as a second language. The T1 project combines fundamental research from A4 with an authentic setting in order to advance meaning and form assessment in the cotext of feedback generation. I actively work on all aspects of the system, from frontend development to backend code.

From my work on technical implementations of systems originated a strong interest by me on approaches for collecting, representing, and processing multi-faceted linguistic data produced by language learners. I'm especially enthusiastic about statistical approaches to NLP, formal representations such as in UIMA, feedback generation, and multi-layer web application development for example with the Google Web Toolkit.

I am part of the ICALL Research Group.
Copyright © 2018 Björn Rudzewitz. All rights reserved.