Context-Aware Information Reduction
Giorgio Orsi
Info
|
Date |
8th November 2011 (week 5, Michaelmas Term 2011) |
|
Time |
11:30 |
|
Place |
Room 147 |
Abstract
The proliferation of data-intensive websites and the growing availability of pervasive and mobile applications enabled
ubiquitous access to multiple sources of "potential" information. This information often remains "potential" as it is hidden
in distributed and hybrid data sources. On the other hand, data typically require reshaping and integration in order to
be adapted to different contexts of use.
Being able to structure and integrate data does not necessarily correspond
to the ability of interpreting them correctly to obtain valuable information. Information is, in fact, (i) noisy, and should
be appropriately focused to the current application situation and (ii) vast and should be appropriately reduced and shaped
before presenting it to the user.
In this talk we investigate the problem of context-aware information reduction,
which uses context meta-data to reduce and personalise the information space in order to improve data management and, in
particular, query answering. The presented techniques have been exploited by various applications such as decision support
for medical-emergency situations and pervasive, user-centred advertising with promising results.
Further info
|
Related series |
|
