University of Rome “La Sapienza”
University of Rome “La Sapienza”
In the last few years, social annotation through the World Wide Web has shifted from its pioneering early stage to its current technological maturity: popular collaborative tagging communities are attended for many purposes by millions of users worldwide. The TAGora FET EU research project, among a wider community of scholars, has closely accompanied this evolution, by analyzing several collaborative networks born at a high rate during this time. Through deep statistical analysis, many common regularities have been uncovered throughout such systems. As a result, features found in such communities are now considered somewhat “universal”, telling researchers about the underlying social structure and semantics.
Moreover, the research groups involved in TAGora have directly impacted on such evolution, by conceiving and implementing novel projects and algorithms to enhance the power of online collaborative tagging. Websites such as Bibsonomy.org or applications such as MyTag and Tagster embody the wide knowledge developed within TAGora, taking advantage of the theoretical investigation and translating it into practical projects.
However, most of the theoretical work has not fully deployed its applicative potential yet. The study of relevance evaluation techniques, of multiple folksonomy integration and of the sociology of social tagging are still open to scientists’ contribution and the object of ongoing research. Accordingly, the workshop will consist in invited contributions reviewing the work already accomplished within TAGora by the authors themselves and possibly a small number of invited speakers, to disseminate to the general public the main results achieved so far and discuss newborn methods and future applications in such field with the community.
Lectures will last 20 minutes + 10 minutes for questions and discussion
10.30-10.45 Vittorio Loreto (ISI, University “Sapienza” of Rome, Italy)
“An introduction and an Overview: 3 years of tagging”
10.45-11.15 Andrea Baldassarri (University “Sapienza” of Rome, Italy)
“Minimal Modelling of Folksonomies” pdf
The enormous increase of popularity and use of the WWW has led in the recent years to important changes in the ways people communicate. An interesting example of this fact is provided by the now very popular social annotation systems, through which users annotate resources (such as web pages or digital photographs) with text keywords dubbed tags. Understanding the rich emerging structures resulting from the uncoordinated actions of users calls for an interdisciplinary effort. In particular concepts borrowed from statistical physics, such as random walks, and the complex networks framework, can effectively contribute to the mathematical modeling of social annotation systems. Here we show that the process of social annotation can be seen as a collective but uncoordinated exploration of an underlying semantic space, pictured as a graph, through a series of random walks. This modeling framework reproduces several aspects, so far unexplained, of social annotation, among which the peculiar growth of
the size of the vocabulary used by the community and its complex network structure that represents an externalization of semantic
structures grounded in cognition and typically hard to access.
11.15-11.45 Andreas Hotho (University of Kassel, Germany)
“Tag Recommendation in Theory and Practice” pdf
In social bookmarking systems, tag recommenders support the user in assigning tags (ie, freely choosable keywords) to resources he is bookmarking. In our system BibSonomy, we have implemented several recommender techniques and a platform for evaluating different recommenders online. In this talk, we will discuss several recommender approaches that we have evaluated on the BibSonomy data and present the platform within the system. We will also summarize findings of the Knowledge Discovery Challenge of ECML PKDD 2008, which had tag recommendations in BibSonomy as topic, and will present this year’s Discovery Challenge which is on the same topic.
11.45-12.00 Coffee Break
12.00-12.30 Barteb Ochab (SONY-CSL)
“Tagging Usages in the Reality” pdf
As Part of the Tagora Project the SONY Computer science lab concentrated on applying tagging concepts on real world applications. For this purpose Peter Hanappe has initiated the “Linking Linke” project which utilised the vast archive of pictures by Armin Linke and allowed visitors to create their own unique image album. The project explores how coherent collections are made using the technologies of social tagging. The second project by CSL is the “NoiseTube”-project. It is a research project about a new participative approach for mapping and tagging noise pollution involving the general public. The idea is to extend the current usage of mobile phones by turning them into noise sensors enabling each citizen to measure and tag his everyday exposure to noise pollution in order to building a collective noise map with a semantic layer.
12.30-13.00 Martin Szomszor (University of Southampton, UK)
“Modelling Users’ Profiles and Interests based on Cross-Folksonomy Analysis” ppt
With the growth of Web2.0, it is becoming increasingly common for users to maintain a presence in more than one social networking site. For example, one might bookmark web pages in Delicious, upload images in Flickr, listen to music in Last.fm, blog in Technorati, etc. The nature of these pursuits naturally leads users to express different aspects of their interests, such as the places one has visited (through Flickr image tagging), the topics one reads about (through the bookmarking of websites), and the bands one listens too. We provide a set of linked data enabled web services to capture and process an individual’s distributed tagging activities, providing reasoning over tagging data (in the form of sense recognition), metadata about tag senses (for disambiguation), and the automatic suggestion of Profiles of Interests.
13.00-13.30 Rabeeh Abbasi (University of Koblenz, Germany)
“Discovering and Exploiting Semantics in Folksonomies”
Resources in Folksonomies are typically annotated very differently. Quite often there is only limited agreement on the choice of tags and users tend to annotate data with a limited vocabulary. The result is sparse data which makes it hard to find particular information. We combine location-based information of photos with tag annotations to recommend tags to the users when they upload their photos. Additionally, we enhance the feature vector representation by including semantic correlations. Finally, we show how the search experience is improved in the MyTag system by interactive word sense disambiguation.
Workshop Organizing Committee
Vittorio Loreto (University of Rome “La Sapienza” – ISI, Turin, Italy)
Steffen Staab (University of Koblenz, Germany)
Harith Alani (University of Southampton, UK)
Gerd Stumme (University of Kassel, Germany)
Luc Steels (Sony Computer Science Lab, Paris, France)
TAGora project started on June 1st 2006
Sixth Framework Programme, Information Society Technologies, IST call 5, Contract N. 34721
Powered by WordPress | Entries and comments feeds | Valid XHTML and CSS