Multi-layer relation networks for relational reasoning

Marius Jahrens, Thomas Martinetz

Abstract

Relational Networks (RN) as introduced by Santoro et al. in 2017 have demonstrated strong relational reasoning capabilities with a rather shallow architecture. Its single-layer design, however, only considers pairs of information objects, making it unsuitable for problems requiring reasoning across a higher number of facts. To overcome this limitation, we propose a multi-layer relation network architecture which enables successive refinements of relational information through multiple layers. We show that the increased depth allows for more complex relational reasoning by applying it to the bAbI 20 QA dataset, solving all 20 tasks with joint training and surpassing the state-of-the-art results.

Original languageEnglish
Title of host publicationAPPIS '19: Proceedings of the 2nd International Conference on Applications of Intelligent Systems
EditorsNicolai Petkov, Nicola Strisciuglio, Carlos M. Travieso
Number of pages5
PublisherACM
Publication date07.01.2019
Pages1–5
Article number10
ISBN (Print)978-1-4503-6085-2
DOIs
Publication statusPublished - 07.01.2019
Event2nd International Conference on Applications of Intelligent Systems - Museo Elder of Science and Technology, Las Palmas de Gran Canaria, Spain
Duration: 07.01.201909.01.2019
Conference number: 149535

Research Areas and Centers

  • Research Area: Intelligent Systems
  • Centers: Center for Artificial Intelligence Luebeck (ZKIL)

DFG Research Classification Scheme

  • 409-05 Interactive and Intelligent Systems, Image and Language Processing, Computer Graphics and Visualisation

Fingerprint

Dive into the research topics of 'Multi-layer relation networks for relational reasoning'. Together they form a unique fingerprint.

Cite this