May M., Saitta L. (eds.) Ubiquitous Knowledge Discovery. Challenges, Techniques, Applications

Springer, 2010. — 248 p.Over the last years, ubiquitous computing has started to create a new world of small, heterogeneous, and distributed devices that have the ability to sense, to communicate and interact in ad hoc or sensor networks and peer-to-peer systems. These large-scale distributed systems, in many cases, have to interact in real-time with their users. Knowledge discovery in ubiquitous environments (KDubiq) is an emerging area of research at the intersection of the two major challenges of highly distributed and mobile systems and advanced knowledge discovery systems. It aims to provide a unifying framework for systematically investigating the mutual dependencies of otherwise quite unrelated technologies employed in building next-generation intelligent systems: machine learning, data mining, sensor networks, grids, peer-to-peer networks, data stream mining, activity recognition, Web 2.0, privacy, user modeling and others.
In a fully ubiquitous setting, the learning typically takes place in situ, inside the small devices. Its characteristics are quite different from currently mainstream data mining and machine learning. Instead of offline-learning in a batch setting, sequential learning, anytime learning, real-time learning, online learning, etc.—under real-time constraints from ubiquitous and distributed data— is needed. Instead of learning from stationary distributions, concept drift (the change of a distribution over time) is the rule rather than the exception. Instead of large stand-alone workstations, learning takes place in unreliable, highly resource constrained environments in terms of battery power and bandwidth.
To explore this emerging field of research, a networking project has been funded since 2006 by the European Commission under grant IST-FP6-0213211: KDubiq (knowledge discovery in ubiquitous environments) is a coordination action at the intersection of the two major challenges of highly distributed and mobile systems and advanced knowledge discovery systems. A basic assumption of the project is that what seems to be a bewildering array of different methodologies and approaches for building smart, adaptive, intelligent ubiquitous knowledge discovery systems can be cast into a coherent, integrated set of key areas centered on the notion of learning from experience. The objective of KDubiq is to provide this common perspective, and to shape a new area of research.
For doing so, the KDubiq coordination action has coordinated relevant research done on learning in many subfields, including:
– machine learning and statistics
– knowledge discovery in databases or data mining
– distributed and embedded computing
– mobile computing
– human computer interaction (HCI)
– cognitive science
A major goal was to create for the first time a forum to bring these individual research lines together, to consolidate the results that have already been achieved, and to pave the way for future research and innovative applications. For doing so, KDubiq has organized a large number of workshops, summer schools, tutorials and dissemination events to bring together this new community.2 One important means to focus the activities and discussions was a collaborative effort to provide a blueprint for the design of ubiquitous knowledge discovery systems. A number of working groups on relevant topics have been established. Their goal was to create a conceptual framework for this new line of research, to survey the state of the art, and to identify future challenges, both on the theoretical and the applications side.
The result of this collaborative effort is Part I of this book. This blueprint manifests the vision and serves as a practical guide for further, integrated advances in this field, towards, in the long-term, building truly autonomous intelligent systems.A Blueprint for Ubiquitous Knowledge Discovery
Introduction: The Challenge of Ubiquitous Knowledge Discovery
Ubiquitous Technologies
Resource Aware Distributed Knowledge Discovery
Ubiquitous Data
Privacy and Security in Ubiquitous Knowledge Discovery
A Human-Centric Perspective on Ubiquitous Knowledge Discovery
Application Challenges for Ubiquitous Knowledge Discovery
Case Studies
On-Line Learning: Where Are We So Far?
Change Detection with Kalman Filter and CUSUM
A Geometric Approach to Monitoring Threshold Functions over Distributed Data Streams
Privacy Preserving Spatio-temporal Clustering on Horizontally Partitioned Data.
Nemoz — A Distributed Framework for Collaborative Media Organization
Micro Information Systems and Ubiquitous Knowledge Discovery
MineFleet: The Vehicle Data Stream Mining System for Ubiquitous Environments