Transcript Poster - Department of Computer Science

Fine-Grain Adaptation Using Context Information
Iqbal Mohomed
Department of Computer Science
University of Toronto
Advisor: Prof. Eyal de Lara
Challenge
Web Adaptation: Factors to consider
• Writing distributed applications that work well in
a mobile environment is hard
• Limited resources
• Heterogeneous Devices
• Wireless link type/quality varies with
mobility
•
Usage Context
• Varying Relevance
•
•
Downloaded
Data  600KB
Network
Connectivity
1 Mbps (DSL)
3G (384Kbps)
Time
(seconds)
5
12.5
GPRS (40Kbps)
120
Multiple Usage
Situational Context
Adaptation can help!
Challenge
• How to pick appropriate adaptation?
• Existing techniques based on
rules/constraints do not consider semantics of
content
Thesis
Use context information to determine semantics
of content and adapt based on this information
•
We consider two applications:
• Web Adaptation
• Remote Health Monitoring
• Context such as the device type, wireless link,
and location can impact desired adaptation
• Context that is most relevant may vary based
on the content as well as type of adaptation!
• We address these challenges with the ContextAware URICA (CA-URICA) technique
• Automatically determine the impact of
different contextual characteristics
• Group users into communities based on
significant context
Remote Health Monitoring
Bluetooth,
ZigBee, etc.
One size does not fit all!
•
•
Web Adaptation: Tailoring
Content By Situational Context
Web Adaptation:
Tailoring Content By Usage Context
• Different objects have varying relevance to the
user, which impacts their adaptation requirements!
• We account for these differences using the
Usage-Aware Interactive Content Adaptation
(URICA) technique
• Provide user with a default adaptation,
and allow content to be adapted interactively
• Use history of successful adaptations to
tailor content for future users
Wifi,
GPRS, etc.
• Context-Aware Filtering can significantly reduce
the amount of data transmitted
• Use context information to judge what
sensor readings are expected
• Vary fidelity of transmitted data based on
whether sensor readings conform to
expectations
Conclusions
• Use context information to determine relevance
of data in a given situation
• When resources are constrained, optimize
based on relevance
Collaborators
Web Adaptation @ University of Toronto:
Prof. Eyal de Lara, Jin Zhang, Jim Cai,
Sina Chavoshi and Alvin Chin
Remote Health Monitoring @ IBM T.J. Watson:
Dr. Maria Ebling, William Jerome, Dr. Archan Misra