WeKnowIt is a 3 year Integrated Project funded by the EU FP7 ICT Programme. The project started in April 2008 with the purpose to develop novel techniques for exploiting multiple layers of intelligence from user-generated content, which transform the large-scale and poorly structured Social Media to meaningful topics, entities, points of interest, social connections and events. WeKnowIt has a budget of 7.5M euro, of which 5.37M euro are funded by the EU.
Currently, there are excellent opportunities to improve European competitiveness in the market of emerging intelligence by investing in new technologies, which can analyse and structure user generated content making it useful for various applications including tourism, news, infotainment, art, e-participation, business and market intelligence. This benefits both organisations and end-users directly since they can receive personalised recommendations about content items, events and topics but also about groups and other users sharing similar interests. Recent progress in research and system development has made it possible to fruitfully combine knowledge occurring in different media to an extent that would have been impossible some years ago. Europe has the chance to corner substantial competitive advantage by the innovations which are generated by WeKnowIt.
To what extent could the “WeKnowIt” project pave the way for further innovation in the field of Web 2.0? Could you give us three main findings of your research?
Existing Web 2.0 technologies for information uploading and sharing as embodied in services such as Flickr, YouTube, Twitter, etc. provide exciting new opportunities to create innovative services. However, such approaches have reached a number of important limits in their evolution, showing inability to “understand” user generated content and user’s social activity and, more specifically, inability to manipulate the content automatically, leading to failure in making information available for further processing, and therefore failing to exploit the emergence of trends at the social and mass level, and the emergence of knowledge about a situation.
WeKnowIt is providing the required technology to support a paradigm shift, setting a foundation for a new generation of services and tools that combine different layers of intelligence for analysing the user generated digital content.
For example, the ClustTour application combines a graph-based image clustering algorithm, making use of both visual and tag information of Flickr photos, with a cluster classification and merging module. The result is an online city exploration application that helps users identify interesting spots in a city by use of photo clusters corresponding to landmarks and events.
Another Collective Intelligence approach is the WeKnowIt Image Recognizer, an application that allows mobile phones users to quickly discover the location and name of photographed objects. The technology behind combines visual and mass content analysis. Image recognition and localization is achieved through a content-based image search engine comprising 1 million Flickr images from 30 European cities, while information about the recognized landmarks is gathered from a list of geo-coded Wikipedia pages.
The WeKnowIt recommendation service exploits the natural multi-modality of social media resources uploaded to Web 2.0 sites, analyzes and combines intelligence from different modalities and offers recommendations that are based on a joint weighted use of Media, Mass and Social Intelligence. The proposed service outperforms baseline recommendation techniques that are based on one kind of intelligence alone.
Overseas ICT players take advantage of their size to process data and generate new income. European SMEs suffer from limited R&D investments in mass-generated content analysis and innovative organisational process. That’s where your project comes. How can European innovative SMEs better analyse data to exploit this new network economy?
Even today there is still a trade off between the high interests in Web 2.0 applications on the one hand and the opportunities to commercialise these products on the other hand. As the web becomes increasingly important for private, commercial and public actors, the WeKnowIt paradigm becomes yet more necessary to close this gap by providing tools that support the integration of enhanced techniques to extract meta-information from content sources. Although current Web 2.0 applications allow and are based on annotations and feedback by the users, these are not sufficient for extracting this “hidden” knowledge, because they lack clear semantics and it is the combination of visual, textual and social context, which provides the ingredients for a more thorough understanding of social content.
Can you give us best practices that linked web-content analysis to enterprises’ decision making?
A generalized example of how Collective Intelligence impacts decision making is the WeKnowIt Emergency Response use case. The analyzed data is collected from a number of different sources: ER workers at the scene of an emergency, general public observing an emergency or other parties publishing information on the web. The information is geo-located, either from metadata provided with the images or through an analysis of textual or visual data and tags are generated for the information. ER workers have the ability to define an incident – automatic processing then connects individual uploads to an overall document in order to provide some categorisation of the information. Finally, techniques from the personal intelligence layer are applied to make the information more palatable for presentation to the ER management personnel who are charged with the task of quick and to-the-point decision making.
A network of research hubs and ICT stakeholders supported “WeKnowIt”. How did they take advantage of being part of the project? Are there any immediate spin-offs for stakeholders?
Most academic partners of the project are exploiting the project outcomes through partnerships with small SMEs (targeted at testing and commercializing individual techniques developed in WeKnowIt), or by protecting their IPRs through patent applications. However they also consider exploitation via sharing of knowledge in the appropriate technical and scientific communities. Industrial partners are in the process of embedding WeKnowIt techniques into their products, e.g. Yahoo’s faceted image search module. They are also protecting their IPRs through patent applications.
Moreover, one spin-off company has been established by University of Koblenz-Landau, namely Kreuzverweis.com, which kicked off in September.