Exploring and Navigating:
Tools for GermaNet
Marc Finthammer and Irene Cramer
Faculty of Cultural Studies, Dortmund University of Technology, Germany
1. Motivation 3. GermaNet Pathfinder
• GermaNet1 is regarded to be a valuable resource for • Features:
many German NLP-applications, corpus research and – calculation of semantic relatedness between given word pair or
teaching; synset pair – 11 relatedness measures available: 8 GermaNet-
based, 3 Google-based
• Unfortunately, the user is left on his own with a bunch – visualization of the respective path (Figure 4) and visualization
of XML-files and insufficient means for navigation and of synsets / subtrees / complete hyponym-tree
– automatic calculation of relatedness for a complete list of word
– inspection and extraction of many features of GermaNet and its
• While intensively working with it in our project HyTex2, hyponym-tree representation
we implemented two GUI-based tools for exploration
and navigation – both are freely available for download
the GermaNet Explorer and
the GermaNet Pathfinder
2. GermaNet Explorer
– word sense retrieval function (Figure 1, region 1)
– structured presentation of all semantic relations pointing to/from Figure 3
the synonym set (synset) containing the currently selected word
sense (Figure 1, region 2)
– visual, graph-based navigation function
– representation of all synsets / wordsenses, (Figure 2) as well as
retrieval, filter, and sort functions (Figure 2, region 2)
Figure 1 4. Conclusion & Future Work
• We already successfully utilized the GermaNet Explorer in
various areas of our research and teaching: e.g. in
experiments on the manual annotation of lexical chains and
for the visualization of lexical semantic concepts in courses.
• GermaNet Pathfinder has successfully been used in our
lexical chainer GLexi; we also found it very helpful to explore
GermaNet and to retrace semantically motivated paths.
• We currently adapt both tools to the brand-new GermaNet
Figure 2 Version.
1 Lothar Lemnitzer and Claudia Kunze (2002): Germanet - representation, visualization, application. In Proceedings of the Language Resources and Evaluation Conference (LREC2002).
2 The HyTex (see www.hytex.info) project, funded the German Research Foundation (DFG), aims at developing strategies for the text-to-hypertext-conversion based on text-grammatical
features. Various sub-tasks of our research, such as lexical chaining or the automatic extraction of definitions, highly rely on the semantic knowledge represented in GermaNet.
3 Jiang and Conrath, 1997 / Leacock and Chodorow, 1998 / Lin, 1998 / Resnik, 1995 / Wu and Palmer, 1994 – please also refer to our Paper for detailed information about the
relatedness measures implemented.