Group 1: Mapping (Renee Miller)
Technical challenges associated with integration at the conceptual
Topics include: Data models; uncertainty (in data and mappings); use
of constraints in integration; mapping formalisms- expressiveness
(declarative, procedural, and trigger-based schemes); schema
evolution; directionality/trust (unidirectional vs. bidirectional or
invertible); interactions with constraints; "combining modalities"
(mappings over text; unstructured data; semi-structured data;
structured data); using mappings in data transformation, schema
evolution, data integration (answering queries using views), data
exchange, querying, search, and ranking in integrated systems;
manipulating/managing mappings (mapping composition/inversion,
integrated view maintenance-notification, subscription, mapping
maintenance and adaptation).
Group 2: Matching (David
Technical challenges associated with integration at the instance
Topics include: Instance matching; interactions of instance with
schema matching; normalization and ontologies; uncertainty in
matching and data; reasoning with uncertainty; probabilistic joins;
using contraints to clean and integrate data; data quality; "hands
off" integration and automated matching; keyword-based search.
Group 3: Integration Architectures (H.V. Jagadish)
"Modes" of integration (views, manual copy-paste, "transclusion",
mash-ups); "end-to-end" integration (assemblies of II systems); how
communities establish the will to collaborate and a common
understanding (schema, ontology) of the sources they wish to
integrate; designing/preparing sources for integration; important
"non-technical" issues: social, legal, economic, etc.
Group 4: Information Integration in Context (Peter Buneman)
This session will look at related technical issues and the context
in which we use II.
Data quality (or selecting the "best" source based on
availability/performance rather that quality); source discovery;
source enhancement; quality assessment and provenance; conveying
annotation and provenance; integrated view maintenance; data
cleaning; security and privacy.