Paper: General estimation and evaluation of compositional distributional semantic models

ACL ID W13-3206
Title General estimation and evaluation of compositional distributional semantic models
Venue Continuous Vector Space Models and their Compositionality
Session
Year 2013
Authors

In recent years, there has been widespread interest in compositional distributional semantic models (cDSMs), that derive meaning representations for phrases from their parts. We present an evaluation of al- ternative cDSMs under truly comparable conditions. In particular, we extend the idea of Baroni and Zamparelli (2010) and Guevara (2010) to use corpus-extracted examples of the target phrases for param- eter estimation to the other models pro- posed in the literature, so that all models can be tested under the same training con- ditions. The linguistically motivated func- tional model of Baroni and Zamparelli (2010) and Coecke et al. (2010) emerges as the winner in all our tests.