Research Areas

Educational Natural Language Processing

Educational Natural Language Processing (Educational NLP or e-NLP) aims at

• finding new applications of Natural Language Processing for educational purposes, and
• developing new techniques and software taking into account the specific needs in the educational domain.

We especially focus on NLP applications for eLearning 2.0, which is characterized by a worldwide learning community where educational material is produced both by students and teachers. This brings about new challenges for NLP since the amount of user-generated discourse and social media content such as wikis and blogs is constantly growing and requires intelligent automatic processing.

Language Technologies for eHumanities

The Language Technologies for eHumanities research group aims at

• research of novel language technology methods to support social sciences and humanities research, and
• scalable and user-friendly web-based services for Digital Humanities.

These technologies are developed to support the imminent paradigm change in humanities and social sciences from small, individual studies to answering (interdisciplinary) research questions supported by a large empirical basis. The group is also interested in software engineering approaches and best practices to build a highly modular, sustainable, and open-source natural language processing framework known as the Darmstadt Knowledge Processing Software Repository (DKPro).

Natural Language Processing and Wikis

Natural Language Processing and Wikis (NLP and Wikis) is a twofold area of research:

• In Wikis4NLP, we investigate how NLP algorithms can be improved by leveraging the vast amount of implicit knowledge prevalent in today's wikis rather than by using traditional knowledge sources like WordNet. Our Application Programming Interfaces for both Wikipedia (JWPL) and Wiktionary (JWKTL) allow for an efficient access to a large pool of collaboratively constructed knowledge.
• In NLP4Wikis, we address the problem that wikis tend to get disorganized as result of their rapid growth. Using NLP techniques, we try to improve user interaction by providing suggestions and hints for everyday tasks. This work is done in the context of the project Wikulu – Self-Organizing Wikis.

Semantic Information Management (SIM) leverages semantic processing techniques for adding structure to unstructured information for more accurate, high-precision and high-recall information search and retrieval.

Information comes in various forms and formats, including business documents, web pages, user manuals, FAQs, and software documentation. In an ever increasing mass of information, finding the right piece of information is becoming more and more difficult.

Statistical Semantics

The Statistical Semantics research group examines statistical methods that reflect natural-language semantics. Specifically we compute semantic similarities and semantic relations between lexical items through the analysis of large texts. These relations are used in applications such as semantic indexing, paraphrasing, and identification of lexical chains.

With a focus on unsupervised and knowledge-free methods, the group advocates the use of crowdsourcing for validation and the creation of training data. In a current project, topic models are employed for modeling lexical cohesion.