Katja Filippova

Brandschenkestrasse 110
8002 Zurich
email: katjaf at search engine
BLR, Dec'10

Research Interests

  • text-to-text generation and summarization
  • information extraction
  • discourse coherence and information structure
  • cognitive linguistics and language evolution

I did my PhD at EML Research (now HITS) while being a student at TU Darmstadt. My thesis is about sentence fusion, that is about how to generate a single novel sentence from the dependency structures of a set of related sentences. I spent three months at Yahoo! Research in Barcelona in 2008, working on how summarization algorithms can be adapted to the domain of financial news. In 2007 I spent half a year at Xerox Research in Grenoble working on structured pdf-to-xml conversion.


  • Modelling Events through Memory-based, Open-IE Patterns for Abstractive Summarization | (talk at ACL 14). Baltimore, MD, June 22-27, 2014. (PDF)
    Daniele Pighin, Marco Cornolti, Enrique Alfonseca and Katja Filippova
  • Opinion Mining on YouTube | (poster at ACL 14). Baltimore, MD, June 22-27, 2014. (PDF)
    Aliaksei Severyn, Olga Uryupina, Barbara Plank, Alessandro Moschitti and Katja Filippova
  • Overcoming the Lack of Parallel Data in Sentence Compression | (talk at EMNLP 13). Seattle, WA, October 18-21, 2013. (PDF)
    Katja Filippova and Yasemin Altun
  • User demographics and language in an implicit social network | (talk at EMNLP 12). Jeju, Korea, July 12-14, 2012. (PDF)
    Katja Filippova
  • Pattern learning for relation extraction with hierarchical topic models | (short paper at ACL 12). Jeju, Korea, July 8-14, 2012. (PDF)
    Enrique Alfonseca, Katja Filippova, Jean-Yves Delort and Guillermo Garrido
  • Improved video categorization from text metadata and user comments | (talk at SIGIR 11). Beijing, China, July 24-28, 2011, to appear. (PDF)
    Katja Filippova and Keith B. Hall
  • Multi-sentence compression: Finding shortest paths in word graphs | (talk at COLING 10). Beijing, China, August 23-27, 2010, pp. 322-330. (PDF)
    Katja Filippova
  • Tree Linearization in English: Improving Language Model Based Approaches | (short paper at NAACL-HLT 09, short). Boulder, Colorado, May 30 - June 5, 2009, pp. 225-228. (PDF)
    Katja Filippova and Michael Strube
  • Company Oriented Extractive Summarization of Financial News | (talk at EACL 09). Athens, Greece, March 30 - April 3, 2009, pp. 246-254. (PDF)
    Katja Filippova, Mihai Surdeanu, Massimiliano Ciaramita and Hugo Zaragoza
  • Sentence Fusion via Dependency Graph Compression | (talk at EMNLP 08). Honolulu, Hawaii, October 25-27, 2008, pp. 177-185. (PDF)
    Katja Filippova and Michael Strube
  • Dependency Tree Based Sentence Compression | (talk at INLG 08). Salt Fork, Ohio, June 12-14, pp. 25-32. (PDF)
    Katja Filippova and Michael Strube
  • German Vorfeld and Local Coherence | Special Issue on Coherence in Dialogue and Generation of Journal of Logic, Language, and Information (JoLLI). Volume 16(4), pp. 465-485.
    Katja Filippova and Michael Strube
  • Generating Constituent Order in German Clauses | (talk at ACL 07). Prague, Czech Republic, June 23-25, 2007, pp. 320-327. (PDF)
    Katja Filippova and Michael Strube
  • Extending the Entity-grid Coherence Model to Semantically Related Entities | (poster at ENLG 07). Schloss Dagstuhl, Germany, June 17-20, 2007, pp. 139-142. (PDF)
    Katja Filippova and Michael Strube

About Me

I am from St. Petersburg, Russia, an incredibly interesting and beautiful city which is getting less and less beautiful thanks to governmental efforts. For (non-)standard pictures check this LJ community where photos are posted daily. Other places I've lived in are Tübingen, Heidelberg, Grenoble and Barcelona.


If you want to learn Russian, I highly recommend the Privet! (Привет!) school in Kaliningrad: the teachers there are very professional and have a lot of experience. The school organizes excursions over weekends, so you won't get bored. Also, Kaliningrad is next to the Baltic sea and is worth visiting on its own!