To the top

Page Manager: Webmaster
Last update: 9/11/2012 3:13 PM

Tell a friend about this page
Print version

Pattern recognition is no… - University of Gothenburg, Sweden Till startsida
Sitemap
To content Read more about how we use cookies on gu.se

Pattern recognition is not enough: representing language, action and perception with modular neural networks

Conference paper
Authors Simon Dobnik
John D. Kelleher
Published in Proceedings of LondonLogue - Semdial 2019: The 23rd Workshop on the Semantics and Pragmatics of Dialogue
ISSN 2308-2275
Publisher Queen Mary University of London
Place of publication London, UK
Publication year 2019
Published at Department of Philosophy, Linguistics and Theory of Science
Language en
Links semdial.org/anthology/papers/Z/Z19/...
https://semdial2019.github.io/#
https://gup.ub.gu.se/file/207840
https://gup.ub.gu.se/file/207867
Keywords spatial descriptions, deep learning, neural networks, modular architectures, language and vision, image captioning
Subject categories Computational linguistics, Linguistics, Cognitive science

Abstract

Current deep learning approaches to modelling of spatial language in generating image captions have shortcomings because they are focused on recognition of visual patterns. The multiplicity of factors that influence spatial language which also include aspects of interaction between speakers and between speakers and their environment invites a modular approach where the solution can be built in a piece-wise manner and then integrated. We call this approach where deep learning is assisted with domain knowledge expressed as modules that are trained on data a top-down or mechanistic approach to otherwise a bottom-up phenomenological approach.

Page Manager: Webmaster|Last update: 9/11/2012
Share:

The University of Gothenburg uses cookies to provide you with the best possible user experience. By continuing on this website, you approve of our use of cookies.  What are cookies?