Collaborative model acquisition

Karl Aberer, EPFL (EPFL Lausanne, Switzerland)

CSIRO ICT

DATE: 2013-03-25
TIME: 14:00:00 - 15:00:00
LOCATION: Seminar room S206 Level 2 (CSIRO ICT Centre, Acton)
CONTACT: JavaScript must be enabled to display this email address.

ABSTRACT:
We observe that many approaches in environmental monitoring, such as participatory sensing, community sensing, opportunistic sensing etc., are based on collaborative techniques for obtaining measurement data. The ultimate goal of environmental monitoring systems is to create from such data models that describe a phenomenon with sufficient accuracy. We also see that very similar approaches are taken in many Web-based applications where collaborative techniques are used to establish a model of some Web-based phenomenon. Examples are recommender systems, reputation systems, collaborative searches systems and many others.

We propose that there exists a class of systems that are focused on collaboratively creating a model of some real phenomenon. These systems are more specific than general collaboration systems, but also more general then domain-specific systems such as those for environmental monitoring. We call the problem these systems tackle collaborative model acquisition.

In this talk we will review different approaches to exhibit principles that we consider as central for collaborative model acquisition. We will first present two approaches from the domain of environmental monitoring illustrating the principles of model cover and utility-driven data acquisition. We then demonstrate in the domain of distributed data integration the use of crowd-sourcing techniques and how to assess the quality of contributions of different agents. We finally turn our attention to the problem of creating systems that make collaboration a rational behaviour using reputation mechanisms. These can be applied in any settings where autonomous agents are supposed to provide services.

Taking such a more generic perspective on the problem of collaborative model acquisition we hope that different communities and disciplines will be able to mutually benefit from their results and techniques across specialized application domains.

BIO:
Karl Aberer is a full professor for Distributed Information Systems at EPFL Lausanne, Switzerland, since 2000. From 2005 to 2102 he is the director of the Swiss National Research Center for Mobile Information and Communication Systems (NCCR-MICS, www.mics.ch). Since September 2012 he is vice-president of EPFL, in charge of information systems. Prior to his current position, he was senior researcher at the Integrated Publication and Information Systems institute (IPSI) of GMD in Germany. He received his Ph.D. in mathematics in 1991 from the ETH ZArich. His research interests are on semantics and self- organization in information systems with applications in peer-to-peer search, semantic web, trust management and mobile and sensor networks. He is or has been serving on the editorial boards of SIGMOD Record, VLDB Journal, ACM Transaction on Autonomous and Adaptive Systems and WorldWide Web Journal and been co-chairing among others the ICDE, ISWC, MDM, ODBASE, VLDB and WISE conferences. Web: http://lsir.epfl.ch/aberer