July 6, 2008, Reykjavik, Iceland
1-day workshop to be held in conjunction with ICALP 2008.
Organizers: Stefano Leonardi (Sapienza University of Rome) and Friedhelm Meyer auf der Heide (Heinz Nixdorf Institute - University of Paderborn)Confirmed Speakers:
09:00 - 09:45 Ashish Goel: Incentives based robust reputation and recommendation systems
09:45 - 10:30 Maurizio Lenzerini: Integration and filtering of heterogeneous information
10:30-11:00 Coffee break
11:00 - 11:45 Muthukrishnan: Streaming and sampling algorithms for information aggregation in networks
11:45 - 12:30 Giuseppe Persiano: Security of networks of low capability devices
12:30-14:00 Lunch break
14:00 - 14:45 Christian Scheideler: Algorithms for scalable and robust information systems
14:45 - 15:30 Open discussion with the participation of the speakers.
15:30-16:00 Coffee break
Network information management is a growing area of research at the attention of several communities: Algorithms, Formal Methods, Information systems, Data mining, Networks and Distributed Systems. The reason of this interest is given by the recent emergence of information systems distributed across the network, pervasive systems, physical mobile and virtual networks of users. These systems are often self-organized, interconnected, highly dynamic, expand up to include the users as source of content and intelligence, ask for a unifying view of data, services and users.
Query, search and retrieval of structured and unstructured information in such systems has already prompted a large body of research at the intersection between databases, information retrieval and networks. Examples are peer-to-peer systems, distributed hash tables, distributed information retrieval, filtering and integration of information, representation, query processing and compression of structured data. Computing aggregate information and statistics in these systems require novel sampling, hashing and streaming techniques that operate with limited storage and computational resources and/or by looking only at a suitable subset of the data. In pervasive systems and networks of sensing entities, local statistics often need to be computed at several vantage points in the network with even more severe restrictions.
These network systems are open and thus prone to manipulation of malicious individuals and collectives. Mechanisms that protect the system against manipulation using incentives, reward and penalties are a wide subject of study as well as the equilibria they induce. On the other hand, the devise of trust and reputation mechanisms that cannot be manipulated by users and coalitions of users are also an important field of investigation. We also observe a growing role of the players in performing fundamental tasks of the network. Users provide tags, annotation, feedback in terms of preferences and opinions. Modelling user behaviour and learning about user profiles from logs and past behavior is therefore becoming crucial in several applications.
These challenges ask for a more solid asset of the theoretical foundations of information management in networks. A number of position papers and recent results will be presented by key researchers in the area that will be invited to this workshop. Several of these subjects are already of interest of the ICALP community, all tracks. However, we believe that a more focussed and thorough presentation of the main challenges of these areas will make an original contribution to the conference.
Click here for information about how to register for this workshop. It is not necessary to register for the main ICALP conference in order to participate in this event.