Kolloquiumsplan FS 2016

Kolloquium FS 2016: Berichte aus der aktuellen Forschung am Institut, BA/MA-Arbeiten, Programmierprojekte, Gastvorträge

Zeit/Ort: Circa alle 14 Tage am Dienstag von 10.15 Uhr bis 12.00 Uhr in Raum BIN 2.A.10.

Dozierende: Annette Rios und Martin Volk

Kontakt: Annette Rios

Datum

Vortragende / Thema

23. Februar 2016

- Volker Dellwo (Phonetisches Laboratorium, UZH): Verbrecher mit ihrer Sprache überführen: Was kann Phonetik und Linguistik leisten?

 

08. März 2016

- Gastvortrag: Sebastian Padó (Stuttgart): Einzel- und mehrsprachige distributionelle Modellierung1

22. März 2016

- Gastvortrag: Jan Šnaider (Zagreb): DErivBase: Derivational Morphology meets Distributional Semantics2

05. April 2016

- Peter Makarov: Towards automated protest event analysis: Mining for protest events

- Gastvortrag: Marco Basaldella (Udine): Keyword Extraction

19. April 2016

- Natalia Korchagina: Gold Standard for Temporal Entity Extraction from Medieval German Texts

- Manfred Klenner, Simon Clematide, Anne Göhring: Sentimentanalyse (Eurospider-Projekt)

03. Mai 2016

- Michi Amsler: Legitimitäts-Framing / Neologismen im politischen Diskurs

- Reto Baumgartner: Morphologieanalyse und Lemmatisierung für Schweizerdeutsch (Masterarbeit)

17. Mai 2016

- Mark Fishel: Bayesian Inference and Machine Translation

- Kurzpräsentationen Programmierprojekte:

   - Ann-Sophie Gnehm: Projekt Korpus Stellenmarkt-Monitor

   - Simon Wegmüller: Automatische Klassifikation grosser Dokumentensammlungen

31. Mai 2016

- Mateusz Dolata: Communication Quality in IT-supported Advisory Services

- Kyoko Sugisaki: Multi-Lingual and Multi-Format Text Processing for International Treaties (Projekt SNIS)

1Sebastian Padó (Stuttgart): Einzel- und mehrsprachige distributionelle Modellierung: Siehe http://www.nlpado.de/~sebastian/pub/papers/tacl14_utt.shtml

2 Jan Šnaider (Zagreb): DErivBase: Derivational Morphology meets Distributional Semantics:

Morphological processing is generally recognized as an important step for many NLP tasks. Most work on computational morphology has focused on inflectional morphology, while derivational morphology, concerned with the formation of new words from existing ones (e.g., read → read+er), has received far less attention. Derivation also introduces a fairly specific semantic shift (+er: agentivization), which makes it very interesting from a lexico-semantic perspective. 

In this talk I will present our work on modeling derivational morphology and linking it to distributional semantics. In the first part, I will describe a rule-based framework for inducing so-called derivational families and its application to create a high-coverage German morphological resource called DErivBase. In the second part, I will focus on the lexico-semantic aspects of derivation and present two studies that combine, in different ways, derivational knowledge and distributional semantics. The first study is about "derivational smoothing" of syntax-based vector spaces, in which we use derivational knowledge to improve distributional similarity predictions. The second study goes in the other direction and uses distributional knowledge alongside structural information about derivational rules to perform a "semantic validation" of the DErivBase resource.