IUI 2008: Integrating Rich User Feedback Into Intelligent User Interfaces

Authors: Simone Stumpf, Erin Sullivan, Erin Fitzhenry,
Ian Oberst, Weng-Keen Wong, Margaret Burnett

SUMMARY:
This paper studies a machine learning system the authors have created.  The authors claim the machine needs to be able to absorb keyword rich information from the user.  In order to do this the authors rely on feedback from the user.  In their experiment the authors want to see how user feedback affects machine learning as well as the user experiment.  Ultimately, the authors found that user feedback was reliable because when users provided useful feedback it made the program more accurate.

DISCUSSION:
While this experiment was a success for the authors, I have a few doubts about the entire concept.  The authors main goal seem to be to create an intelligent program with the ability to learn.  If user feedback is needed to make the program smarter, how intelligent is the program?  The main point of a program that can learn is to take work off the user.  If the users are always concerned about giving feedback it takes away from the time they are using the program for intended purposes.