Paper Reading #20: Rush: Repeated Recommendations on Mobile Devices

Comments

Reference

Dominikus Baur, Sebastian Boring, and Andreas Butz

Rush: Repeated Recommendations on Mobile Devices
Intelligent User Interfaces

Summary

This paper talks about a recommendation based interface for mobile devices. While recommendation interfaces are very popular on desktop products, little has been done in order to adapt them to mobile devices. Even when these interfaces have been ported to such devices, they do not take into account orientation of the screen or occlusion problems that occur while controlling the touch interface in such a small device. In a sense, this interface is very similar to Dasher since it uses continuous gestures to increase efficiency.

Discussion

Final project proposal

For the final project, I am working with Kevin Casey in implementing the TA interface that helps then mark students progress in class

http://kcasey436.blogspot.com/2011/04/final-project-proposal.html

Paper Reading #19: Tell Me More, not just "More of the Same"

Comments

Reference
Francisco Iacobelli, Larry Birnbaum, and Kristian J. Hammond
Tell Me More, not just "More of the Same"
Intelligent User Interfaces

Summary

Given the amount of information that is currently available on the internet, it is important that researchers devise better ways of presenting this information to the user. This is particularly true in the context of media outlets and search engines; who increase the amount of information presented by suggesting similar pages to the one requested. It is important to state that it is up to the user to determine whether the suggested articles constitute new information of not. Tell Me More is an interface that can make this classification automatically. Given a seed article, this interface is able to collect new information based on additional quotes, actors, figures, or information. Furthermore, this information is presented with explicit categorization; as opposed to the binary criteria (new/not new) currently used by other systems. The current system implementation consists of five subsystems that help Tell Me More classify and display new information: content gathering, content filtering, text analysis, difference metrics, and presentation. By using LSA (Latent Semantic Analysis), Tell Me More is able to collect a list of different entities (subjects), quantifiers, and quotes in order to determine if the collected information is new and whether it really constitutes new information.

Discussion
This was a very interesting paper in particular because current search engines fail to recognize whether the related material (or suggested readings) represent additional information or not. This is not only seen in the news outlets, but also when shopping for particular products. For example buying a CD that consists of the best music of an artist where several tracks on the disc are already on another disc belonging to the user. I really see how this system can be useful.