Database Reference
In-Depth Information
Data Integration in a Networked World
Karl Aberer
School of Computer and Communication Sciences
Ecole Polytechnique Fed´eraledeLausanne(EPFL)
karl.aberer@epfl.ch
Abstract.
Traditionally, data integration techniques involve central
components, e.g., global schemas or ontologies, to overcome semantic het-
erogeneity for enabling transparent access to heterogeneous data sources.
Today, however, with the explosion of machine processable formats in the
Data Web, one cannot rely on global, centralized schemas anymore, as
knowledge creation and consumption are getting more and more dynamic
and decentralized.
Peer-to-peer data integration systems are a good example of this new
breed of systems eliminating central semantic components and replac-
ing them through decentralized processes of local schema alignment and
query processing. As a result semantic interoperability becomes an emer-
gent property of the system.
In this talk we will first survey recent developments in peer-to-peer
data integration, illustrating which novel challenges and opportunities
these systems introduce. We then illustrate of how semantic integra-
tion can be modeled as a self-organizing agreement process and present
a probabilistic technique for distributed reasoning about semantic rela-
tionships. We show how to extend these methods from the problem of
schema mapping to entity matching and how to consider trustworthi-
ness of autonomous participants in this process. Finally we present some
recent work on applying these principles in business applications.
