|
Group A: Next Generation Information Access Group
Mission Statement
1-Stop Information Shopping by Content
Access across all domains (facilitated by API's)
Scientific
Humanities (music, art, social science)
In all languages and all forms of expression
E.g., Braille, sign language
Research Issues: Universal Data Access
Intuitive access
Web-based, distributed IR
Integration of multimedia/multimodal data types
Software architectures (API, experimental prototypes)
Cross-lingual
DB for XML
Indexing of new data types
Similarity queries
Weighting
Continuum of Correctness
Approximate answer (versus speed)
IR degree of confidence
NLP and Statistics
Incomplete and inconsistent information
New Query Types
Data mining on multimedia data
Data mining as database querying
Rule induction
Association rules
Universal Knowledge Operators (Fundamental) Generalized Access Framework
Representations
Image-Continuous Media
Text
Numbers
Operations
Classification
Clustering
Temporal
Association rules
Rule induction
Integration of Knowledge Representation
NLP and IR (case based reasoning in the large)
Prior knowledge
Management of discovered knowledge
User-centered
Better content representations than "bag of words"
Needs / Enablers
Test Collections and Procedures Depository
Multidimensional, multimedia
Spatiotemporal, deductive, active
Data mining, distributed searching (But has to be done carefully!)
Leisure dbs, photo dbs (zoom angle)
SBIR + IDM coordination?
Software tools?
NSF should advertise success stories |