Title: ViewpointS: capturing formal data and informal contributions into an adaptive knowledge graph

Authors: Philippe Lemoisson; Guillaume Surroca; Clement Jonquet; Stefano A. Cerri

Addresses: TETIS, Univ Montpellier, AgroParisTech, CIRAD, CNRS, IRSTEA, Montpellier, France; CIRAD, UMR TETIS, F-34398 Montpellier, France ' Laboratory of Informatics Robotics Microelectronics of Montpellier (LIRMM), University of Montpellier and CNRS, France ' Laboratory of Informatics Robotics Microelectronics of Montpellier (LIRMM), University of Montpellier and CNRS, France ' Laboratory of Informatics Robotics Microelectronics of Montpellier (LIRMM), University of Montpellier and CNRS, France

Abstract: Formal data is supported by means of specific languages from which the syntax and semantics have to be mastered, which represents an obstacle for collective intelligence. In contrast, informal knowledge relies on weak/ambiguous contributions e.g., I like. Reconciling the two forms of knowledge is a big challenge. We propose a brain-inspired knowledge representation approach called ViewpointS where formal data and informal contributions are merged into an adaptive knowledge graph which is then topologically, rather than logically, explored and assessed. We firstly illustrate within a mock-up simulation, where the hypothesis of knowledge emerging from preference dissemination is positively tested. Then we use a real-life web dataset (MovieLens) that mixes formal data about movies with user ratings. Our results show that ViewpointS is a relevant, generic and powerful innovative approach to capture and reconcile formal and informal knowledge and enable collective intelligence.

Keywords: knowledge representation; knowledge graph; semantic web; social web; collective intelligence; nature-inspired computational model; formal/informal knowledge; serendipitous learning.

DOI: 10.1504/IJKL.2018.092047

International Journal of Knowledge and Learning, 2018 Vol.12 No.2, pp.119 - 145

Received: 07 Oct 2016
Accepted: 20 Sep 2017

Published online: 30 May 2018 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article