On inductive abilities of latent factor models for relational learning

DOI10.1613/JAIR.1.11305zbMATH Open1451.68278arXiv1709.05666OpenAlexW2962834633WikidataQ125761043 ScholiaQ125761043MaRDI QIDQ4611399FDOQ4611399

Authors: Théo Trouillon, Éric Gaussier, Christopher R. Dance, G. Bouchard 

Publication date: 18 January 2019

Published in: Journal of Artificial Intelligence Research (Search for Journal in Brave)

Abstract: Latent factor models are increasingly popular for modeling multi-relational knowledge graphs. By their vectorial nature, it is not only hard to interpret why this class of models works so well, but also to understand where they fail and how they might be improved. We conduct an experimental survey of state-of-the-art models, not towards a purely comparative end, but as a means to get insight about their inductive abilities. To assess the strengths and weaknesses of each model, we create simple tasks that exhibit first, atomic properties of binary relations, and then, common inter-relational inference through synthetic genealogies. Based on these experimental results, we propose new research directions to improve on existing models.

Full work available at URL: https://arxiv.org/abs/1709.05666

Recommendations

• scientific article; zbMATH DE number 7049742
• Reference classes and relational learning
• Demystifying relational latent representations
• Multi-relational learning via hierarchical nonparametric Bayesian collective matrix factorization
• Knowledge graph completion via complex tensor factorization

zbMATH Keywords

knowledge graphsinductive abilityinter-relational inferencesynthetic genealogies

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05) Research exposition (monographs, survey articles) pertaining to computer science (68-02) Knowledge representation (68T30)

Cited In (4)

• Demystifying relational latent representations
• Title not available (Why is that?)
• Reference classes and relational learning
• Leveraging maximum entropy and correlation on latent factors for learning representations

This page was built for publication: On inductive abilities of latent factor models for relational learning

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4611399)