Meanings and meaning shifts of multi-word expressions (MWEs) have been explored by a variety of distributional and multi-modal models, and across many languages. Our own work has focused on English and German noun compounds and particle verbs, by applying and comparing textual and visual information to predict MWE compositionality as well as regularities and analogies in meaning shifts. As a particular perspective in our approaches we balanced gold standard datasets according to MWE and constituent frequencies, productivities and ambiguities, and investigated the diverse models accordingly.
In this talk, I will first present a selection of computational models of noun compound meaning for general language, and then describe domain-specific German term datasets and approaches that model compound term difficulty in interaction with empirical compound and constituent properties and compound--constituent termhood inter-dependencies.
Dr. Sabine Schulte im Walde is associate professor at the Institute for Natural Language Processing at the University of Stuttgart in Germany. She performed her Master studies at the Universities of Stuttgart and Edinburgh and received her PhD in Computational Linguistics in 2003 from the University of Stuttgart and the Venia Legendi (Habilitation) from Saarland University in 2009. From 2003-2004 she worked for the lexicographer Duden in Mannheim, Germany, and from 2011-2016 she was a Heisenberg Fellow. Her work applies statistical methods to lexical-semantic phenomena, with a focus on the linguistic and cognitive plausibility of the computational approaches. The topics of her research include the automatic induction of semantic classifications and semantic relations; compositionality and meaning shifts of multi-word expressions; synchronic and diachronic ambiguity and figurative language usage; and the evaluation of corpus-based semantic knowledge.