Unstructured Information Management

Unstructured Information Management (Project, SS 2012)

Description

Life-long learning is more and more becoming a reality not just a slogan. People learn on their way to work using a smartphone or on the bus or at home on the couch. However, there either is limited learning material which can be used in such situations or it is very expensive. A way to overcome this drawbacks is using Natural Language Processing in order to automatically:

  • create learning materials;
  • augmenting learning materials;
  • generate test items (e.g. multiple-choice or knowledge questions);
  • assess free text answers;
  • or determine the difficulty of test items or texts.

The projects will focus on applications in the learning scenario which can benefit from Natural Language Processing. In each project, a working prototype for some domain (e.g. language learning, computer science courses, etc.) is going to be implemented. Prototypes might be stand-alone applications or smartphone apps.

Literature

  • Lee, J. ; Seneff, S.: Automatic Generation of Cloze Items for Prepositions. In: Proceedings of INTERSPEECH. Antwerp, Belgium, 2007.
  • Mohler, M. ; Mihalcea, R.: Text-to-text Semantic Similarity for Automatic Short Answer Grading. In: Proceedings of the European Chapter of the Association for Computational Linguistics (EACL 2009). Athens, Greece, 2009

Prerequisites and Preparation

  • Programming in Java

Deliverables

  • Kick-off presentation
  • Documented source code / unit tests
  • Project description
  • Final presentation

Timetable

Introductory sessions will be held during the first three weeks Thursdays (12.04., 19.04., 26.04.) from 15:30 to 18:30 in S2|02 D017.

Brief regular status meetings are (tentatively) planned for Thursdays between 16:00 to 17:00. Actual times may vary depending on the number of participants.

Lecturers

  • Dr. Torsten Zesch