Birte Lönneker-Rodman

I am a visiting researcher within the AI group (FrameNet project) at the International Computer Science Institute, Berkeley, CA (since September 2006). My main research areas include: semantic resources for Natural Language Processing; computational lexicography; computational approaches to metaphor and figurative language in general; and Cognitive Linguistics. A selection of recent and current projects is presented below. Further information about my previous work and activities can be found on my Story Generation web page. There is also a German CV, as of May, 2008. (By the way, you might also know me as Birte Loenneker-Rodman.)

FrameNet and multilinguality

FrameNet is creating an on-line lexical resource for English, based on frame semantics and supported by corpus evidence. In Natural Language Processing, the resource is exploited for shallow semantic tagging or "semantic role labeling". I am researching representation structures and abstraction levels in FrameNet and other lexico-semantic NLP resources. To exemplify the practical consequences for multilingual applications, I am investigating the integration of FrameNet methods and information types into a bilingual online dictionary that I created previously, Online SLO-DE-SLO.

Interrelatedness in semantic resources

Publications

• Birte Lönneker-Rodman (2007a): Multilinguality and FrameNet. ICSI Technical Report TR-07-001, International Computer Science Institute, Berkeley, CA. March 2007. 25 pages.
• Birte Lönneker-Rodman (2007b): "Beyond Syntactic Valence: FrameNet Markup of Example Sentences in a Slovenian-German Online Dictionary." In Jana Levická and Radovan Garabík (eds.): Slovko 2007. Computer Treatment of Slavic and East European Languages. Proceedings of the Fourth International Seminar. Bratislava, Slovakia, 25-27 October 2007. (http://korpus.juls.savba.sk/~slovko/2007/) Tribun, Bratislava 2007. 152-164.

Presentations

• A Bilingual Dictionary with Links to FrameNet. Presentation at the FrameNet meeting, ICSI Berkeley, 30 January 2007.
• Multilinguality and FrameNet. Presentation at the FrameNet meeting, ICSI Berkeley, 13 February 2007.
• Semantic {Frames, Types, Web} and Ontologies. Invited presentation at the CIMPL meeting (Center for the Investigation of Mind, Pictures, and Language), Princeton, NJ. 30 April 2007.
• FrameNet and Ontologies. Presentation at the FrameNet meeting, ICSI Berkeley, 15 May 2007.
• Notes on the Kappa Statistic. Presentation at the FrameNet meeting, ICSI Berkeley, 5 September 2007.

FrameNet Markup in a bilingual dictionary (German-Slovenian)

Slovenian FrameNet

At the International Computer Science Institute, we are adapting the FrameNet annotation tool (FrameNet Desktop) to Slovenian. This involved the creation of various related tools and resources. I populated the lexicon part of the FrameNet database with Slovenian data, collected and preprocessed several sample corpora, produced part-of-speech annotated data, and trained an automatic part-of-speech tagger for Slovenian. The corpora have been tagged and loaded into the FrameNet Desktop where they are ready for annotation with semantic frames.

Screenshot of the Slovenian FrameNet annotation tool

Related tools and resources

• Automatic ambiguous morphological tagger (part-of-speech tagger) for Slovenian
• Manually disambiguated morphological annotation of Slovenian texts (over 14,000 tokens)
• Word-form generator for Slovenian nouns, verbs, and adjectives
• Slovenian word form lexicon containing more than 600,000 word forms of roughly 6,000 lemmas
• Slovenian parameter file for automatic part-of-speech tagging with the TreeTagger

Publications

• Birte Lönneker-Rodman, Collin Baker and Jisup Hong (2008): "The new FrameNet Desktop: A Usage Scenario for Slovenian." In Proceedings of ICGL 2008, the First International Conference on Global Interoperability for Language Resources, 9-11 January 2008, Hong Kong. 147-154.

The Hamburg Metaphor Database

Several years ago, I initiated and created the Hamburg Metaphor Database, an online database of predominantly French and some German metaphors, supported by the Institute for Romance Languages at Hamburg University. The database includes a lexical as well as a conceptual layer; currently, it contains more than 1,600 annotated attestations from corpora. Together with Ernesto De Luca (Magdeburg, Germany), I am now exploring possibilities to convert the Hamburg Metaphor Database into RDF/OWL and to make the product available as an add-on to EuroWordNet.

Hamburg Metaphor Database query results for the source domain explosion

Recent publications

• Birte Lönneker-Rodman (forthcoming): "The Hamburg Metaphor Database. Issues in Resource Creation". To appear in Language Resources and Evaluation.
• Birte Lönneker-Rodman and Astrid Reining (2007): "Corpus-driven Metaphor Harvesting." In Proceedings of the NAACL/HTL 2007 workshop on Computational Approaches to Figurative Language. Rochester, NY. April 26, 2007. 5-12.
• Ernesto W. De Luca and Birte Lönneker-Rodman (2008): "Integrating Metaphor Information into RDF/OWL EuroWordNet". To appear in Proceedings of the 6th International Conference on Language Resources and Evaluation (LREC-2008), Marrakech, Morocco. May 28–30, 2008.

Presentations

• Neue Dimensionen der Metaphernforschung. Institut für Linguistik/Romanistik, Stuttgart, Germany. July 14, 2008.
• New dimensions of metaphor research. FrameNet meeting, ICSI, Berkeley, CA. July 8, 2008.
• Behind the Scenes at the Hamburg Metaphor Database. ROLAP meeting (Research on Language at Princeton), Princeton, NJ. May 1, 2007.
• Metaphor Resource Creation. NTL meeting, ICSI, Berkeley, CA (see also ICSI events). September 7, 2007.

Student supervision

• Co-advisor to Astrid Reining (Hamburg, Germany), 2005-2006. Master's thesis: Corpusanalysetools und die Hamburger Metapherndatenbank. Eine computergestützte Analyse von Metaphern im Europadiskurs der französischen Tagespresse. Institute for Romance Languages, University of Hamburg.
• Co-advisor to Anne Simmank (Hamburg, Germany), 2007-2008 (ongoing). Master's project: Das französische Verfassungsreferendum vom 29. Mai 2005. Untersuchungen zum Metapherngebrauch in deutschen und französischen Printmedien.

Computational approaches to non-literal language

Metaphors and other types of non-literal language are a challenge for Natural Language Processing applications. They have a double-layered semantics: consider, for example, the image evoked by a metaphor, and its actual meaning in context. Ideally, translation (including Machine Translation) preserves both the actual meaning and the image - but sometimes, the target language does not offer any lexicalized expression that serves this purpose. Information Retrieval applications, on the other hand, should work with the contextual meaning only; document classification would be influenced negatively if the underlying images were processed. Researchers in the emerging field of Figurative Language Processing are currently exploring techniques, methods, and representational issues related to the computational treatment of non-literal language. We are hoping so see our results integrated into truly semantic-aware NLP systems of the future.

Translating the non-literal (from an evaluation of Machine Translation)

Publications

• Birte Lönneker-Rodman and Astrid Reining (2007): "Dmitrij Dobrovol'skij and Elisabeth Piirainen: Figurative Language: Cross-Cultural and Cross-linguistic Perspectives." Book review. Cognitive Linguistics 18-1. 121-131.
• Birte Lönneker-Rodman and Behrang Mohit (2008): "Translation of the non-literal: Evidence from aligned bilingual corpora." To appear in Abstract proceedings of the third international conference of the German Cognitive Linguistics Association (DGKL/GCLA-2008), Leipzig, Germany.

Events

• Antonietta Alonge (Perugia, Italy) and Birte Lönneker-Rodman (organizers): Figurative language in WordNets and other lexical resources, and their applications. Panel at the Second Global WordNet Conference (GWC-2004), Brno, Czech Republic. January 2004.
• Antonietta Alonge (Perugia, Italy) and Birte Lönneker-Rodman (organizers): Cognitive-linguistic approaches: What can we gain by computational treatment of data? Theme session at DGKL-06/GCLA-06 (Second International Conference of the German Cognitive Linguistics Association), Munich, Germany. 7 October 2006.

Teaching

• Advanced course on Figurative Language Processing, 19th European Summer School in Logic, Language and Information (ESSLLI 2007), Trinity College, Dublin, Ireland. 6 -10 August 2007.

... and some additional information:

Reviewer Global WordNet Conference 2008 Third International Conference of the German Cognitive Linguistics Association, Leipzig 2008 ICSC 2008 (Second IEEE International Conference on Semantic Computing) Sixth Language Technologies Conference, 2008, Ljubljana, Slovenia German Cognitive Linguistics Association I am a board member of the German Cognitive Linguistics Association. You can join us! When I'm not at ICSI... ... I am probably in the climbing gym, or taking care of our daughter Amaia.

Acknowledgments

I am supported by a DAAD (German Academic Exchange Service) postdoctoral fellowship, September 2006 - November 2008.