Spanish FrameNet and its applications to NLP

Carlos Subirats (subirats <at> icsi (point) berkeley (point) edu)

Autonomous University of Barcelona and International Computer Science Institute

Le 4 mai 2011 à 11 h 30

Salle C-9019, Pavillon Lionel-Groulx

The Spanish FrameNet Project (SFN) is creating an online lexical resource for Spanish ( based on Frame Semantics (Fillmore and Baker 2010, Ruppenhofer et al. 2010) and supported by corpus evidence. The aim of SFN is to document the range of semantic and syntactic combinatory possibilities (valences) of each word of Spanish in each of its senses by semiautomatically annotating example sentences and automatically organizing the annotation results. SFN has also begun to apply Construction Grammar (Fillmore and Lee-Goldman 2011) to specifying the syntactic and semantic characteristics of Spanish grammatical constructions and linking them to semantic frames. The SFN corpus of annotated sentences has been used as a training corpus for Shalmaneser (Erk and Padó 2006), a Frame-Semantics-based statistical program for automatic semantic-role labeling. The SFN project has started a pilot study to convert SFN's database of frames and constructions into the formalism of Embodied Construction Grammar (ECG) (Dodge 2010, Feldman 2006, Bergen et al. 2004). The ECG-based Construction Analyzer (Bryant 2008) uses the translated frames and constructions from SFN to automatically analyze sentences of Spanish by generating a semantic specification, i.e., an integrated semantic and syntactic analysis. ** Spanish FrameNet (SFN) is being developed at the Autonomous University of Barcelona (Spain) and the International Computer Science Institute (Berkeley, CA) in cooperation with the FrameNet Project. SFN is sponsored by Spain's Department of Science and Innovation (FFI2008-0875).

Pour recevoir les annonces hebdomadaires par courriel, visitez

Liste des autres séminaires