SpotTheLink: A Game-Based Approach to the Alignment of Ontologies

SpotTheLink: A Game-Based Approach to the Alignment of Ontologies

Stefan Thaler (STI-Innsbruck, University of Innsbruck, Austria), Elena Simperl (AIFB, Karlsruhe Institute of Technology, Germany), Katharina Siorpaes (STI-Innsbruck, University of Innsbruck, Austria) and Stephan Wölger (STI-Innsbruck, University of Innsbruck, Austria)
DOI: 10.4018/978-1-4666-0894-8.ch003
OnDemand PDF Download:
No Current Special Offers


A multitude of approaches to match, merge, and integrate ontologies, and more recently, to interlink RDF data sets, have been proposed over the past years, making ontology alignment one of the most active and at the same time mature area of research and development in semantic technologies. While advances in the area cannot be contested, it is equally true that full automation of the ontology-alignment process is far from being feasible; human input is often indispensable for the bootstrapping of the underlying methods, and for the validation of the results. The question of acquiring and leveraging such human input remains largely unaddressed, in particular when it comes to the incentives and motivators that are likely to make users invest their valuable time and effort in alignment tasks such as entity interlinking and schema matching, which can be domain-knowledge-intensive, technical, or both. In this chapter, the authors present SpotTheLink, a game whose purpose addresses this challenge, demonstrating how knowledge-intensive tasks in the area of the Semantic Web can be collaboratively solved by a community of non-experts in an entertaining fashion.
Chapter Preview

1. Introduction

A large share of tasks in knowledge engineering crucially rely on human input (Siorpaes and Simperl 2010). This applies in particular to those tasks which are acknowledged to be hardly approachable in a systematic, engineering-driven fashion; and also, though to a lesser extent, to the wide array of (semi-) automatic methods and techniques that have been proposed as an attempt to reduce the costs of knowledge-engineering projects by minimizing the need for human involvement in these projects. In this second case, despite constant progress in improving the performance of the corresponding algorithms and the quality of their results, experiences show that human input is nevertheless required, even if it just for the configuration of the algorithms, the creation of knowledge corpora the algorithms can be trained on, or the validation of (intermediary) algorithm outputs. Examples of such knowledge engineering tasks are numerous, including machine translation, information extraction, text summarization, data integration, multimedia analysis, and conceptual modeling.

One novel approach which proved successful to resolve technical tasks via human computation is based on ‘games with a purpose’ (Van Ahn 2006). The idea behind games with a purpose is simple, but effective; tasks which remain difficult to handle by computers, but which humans seem to tackle easily are hidden behind entertaining, collaborative games targeting not experts, but casual Internet users. By playing a game with a purpose, users are indirectly generating data that can be capitalized to build knowledge corpora required for the training of algorithms, thus providing a powerful example of how human and computational intelligence can be combined to address important, challenging problems. Since the original proposal by Van Ahn in 2006 games with a purpose have been applied to tasks as diverse as image and video annotation,1 genetics,2 natural language processing,3 and knowledge engineering.4

Complete Chapter List

Search this Book: