A formal distributional semantics for cognitively-plausible reference acts
Reference -- the ability to talk *about* things -- is one of the most fundamental functions of language. This function is well accounted for in formal semantics, via the classic notions of extension, intension and truth value. The standard formal apparatus is however difficult to transfer to a realistic and cognitively-plausible speaker setup, where models and denotation functions are not shared amongst individuals. In this talk, I propose a hybrid semantic formalisation drawing from set theory and distributional semantics. The set-theoretic aspects of this formalisation retain part of the standard notion of reference, whilst its distributional aspects account for model learnability and thus, speaker differences. I suggest that a combined semantics affords a range of inferences that reconcile standard ideas of denotation with speaker individuality. I then report on a series of experiments that test the learnability of such a formalisation from raw data, showing results on the acquisition of quantified concept-property pairs, the representation of entity vectors, and the acquisition of quantifiers themselves.
Aurelie is a Marie Skłodowska-Curie fellow at the Universitat Pompeu Fabra in Barcelona. Her research focuses on models of semantics that bridge across formal and distributional representations of meaning. She is particularly interested in underspecified models of quantification and the representation of entities in vector spaces. Recently, she has also worked a lot on language and vision models of quantifiers. She holds a PhD in Natural Language Processing from the University of Cambridge. She has held an Alexander von Humboldt fellowship in Potsdam, and postdoc positions in Cambridge, Stuttgart, and the Centre for Mind/Brain Sciences in Trento.
Add to your calendar
17 November 2017 - Aurelie Herbelot: Seminar
Informatics Forum 4.31/4.33