Emerging, Collective Intelligence for Personal, Organisational and Social Use [chapter]

Sotiris Diplaris, Andreas Sonnenbichler, Tomasz Kaczanowski, Phivos Mylonas, Ansgar Scherp, Maciej Janik, Symeon Papadopoulos, Michael Ovelgoenne, Yiannis Kompatsiaris
2011 Studies in Computational Intelligence  
Existing approaches towards extracting Collective Intelligence usually build upon restricted combinations from the available social media attributes. For example, in [18] geo-locations and tag information are used in order to generate representative city maps. In [19] tags and visual information together with geo-location are used for objects (e.g. monuments) and events extraction. Tags from Flickr images and timestamp information are used in [20] to form a chronologically ordered set of
more » ... hically referenced photos and distinguish locals from tourist travelling. The description of city cores can be derived automatically, by exploiting tag and location information [21] . The approach is able of distinguishing between administrative and vernacular uses of place names, thus avoiding the potential for confusion in the dispatch of emergency services. But besides these combinations, user generated content can be viewed as a rich multi-modal source of information including attributes such as time, favorites and social connections. For example, beyond harnessing content and the surrounding tags or text, limited effort has been made to include the social patterns into the media analysis. The important aspect of fusion of modalities and different sources is currently lacking in existing Collective Intelligence applications. In this chapter novel techniques for exploiting these multiple layers of intelligence from user-contributed content are presented, which together constitute Collective Intelligence, a form of intelligence that emerges from the collaboration and competition among many individuals. The Collective Intelligence technologies to be described were developed in the context of the FP7 EU project WeKnowIt: Emerging, Collective Intelligence for personal, organisational and social use 2 . User contributed content is analysed by integrating research and development in visual content analysis for localisation (Media Intelligence), tag clustering and Wikipedia ontology-based categorization (Mass Intelligence), analyzing social structures and user communities access rights (Social Intelligence) and event representation (Organisation Intelligence). The exploitation of the emerging Collective Intelligence results is showcased in two distinct case studies: an Emergency Response and a Consumers Social Group 3 case study. The chapter structure comprises an overall of nine sections. After this Introduction, the following section details the current state-of-the-art for each intelligence layer. The next four sections describe the developed technologies in each intelligence layer individually, namely the Media, Mass, Social and Organisational Intelligence layers. The seventh section presents the integration of the different technologies within one framework, namely the Integrated Collective Intelligence Framework (ICIF), which is exploited in the Emergency Response and Consumer Social Group use cases. The eighth section describes these two application scenarios where the presented techniques have been used together in order to leverage Collective Intelligence. The section also includes user evaluation results for the developed demonstrators. The last section contains conclusions and possibilities on building on top of the achieved Collective Intelligence results.
doi:10.1007/978-3-642-20344-2_20 fatcat:crgix2grbvfupaht5mp3cjriii