Explainable, Verifiable, Relational Representation Learning from Knowledge Graphs - Naver Labs Europe
loader image

UCL LogoThe seminar run from 11am to 12pm. Please register online

Date: 27th June 2019

Speakers: Pasquale Minervini, Research Associate in Statistical NLP & ML at UCL, London, UK

Abstract: Knowledge Graphs are graph-structured knowledge bases, where knowledge about the world is encoded in the form of relationships between entities. We will discuss some of the work we did in the space of identifying missing links in large-scale Knowledge Graphs using neural link predictors, and how you can incorporate background knowledge – in the form of First-Order Logic rules or constraints – in neural link predictors, to generalise from less data and incorporate formal guarantees in neural models. Finally, we will discuss how to reason jointly learn representations and rules via end-to-end differentiable reasoners.


Ceci correspond à une petite biographie d'environ 200 caractéres