University of Edinburgh
Incrementally Predicting Syntax and Semantics
When processing language, speakers incrementally construct syntactic representations and use these to derive lexical, sentential, and discourse meaning. The syntactic aspects of this processes are well captured by incremental parsers and psycholinguistic models such as surprisal. However, there is very little work on modeling the incremental construction of semantic representations. Here, we show how an existing predictive parsing framework (psycholinguistically motivated tree-adjoining grammar) can be augmented with an incremental semantic role labeler, which then feeds into semantic construction using distributed representations. We discuss how surprisal estimates can be derived in this framework, and present an evaluation both on NLP tasks and against psycholinguistic data.
Joint work with Ioannis Konstas.
Frank Keller is professor of computational cognitive science in the School of Informatics at the University of Edinburgh. His background includes an undergraduate degree from Stuttgart University, a PhD from Edinburgh, and postdoctoral and visiting positions at Saarland University and MIT. His research focuses on how people solve complex tasks such as understanding language or processing visual information. His work combines experimental techniques with computational modeling to investigate reading, sentence comprehension, translation, and language generation, both in isolation and in the context of visual information such as photographs or diagrams. Prof. Keller serves on the management committee of the European Network on Vision and Language, is a member of governing board of the European Association for Computational Linguistics, and recently completed an ERC starting grant in the area of language and vision.