Comments:
TBD
Reference Information:
Title: Real-time interaction with supervised learning
Authors: Rebecca Fiebrink of Princeton University, Princeton, NJ, USA
Presentation Venue: CHI EA '10 Proceedings of the 28th of the international conference extended abstracts on Human factors in computing systems
Summary:
The researcher of this paper is working on a project to determine how the design of certain user interfaces will affect the interaction with machine learning algorithms. By utilizing this approach, the researcher hopes that the artificial intelligence concept of machine learning can be easily integrated into a package that works well for a human user in a real-time environment.
One of the researcher's primary goals is to determine how effective this approach will be for users who do not have a lot of technical knowledge. Through this research, they hope to determine how this will influence the design of HCI interfaces and how it will be perceived and utilized.
Discussion:
Machine learning is an intriguing topic. I find the possibilities envisioned in science fiction literature about the potential capabilities of robots or androids to be fascinating subject. I can see that research done in this field could possibly lead to that. This might be one step on the road to that.
Of course, the argument against creating anthropomorphic automatons on the sole criterion of form factor will come up. Why should be make robots in our own image? I think that many robots do not need to be made up with a human likeness. Robots that would receive greater utility from a different form factor, like having wheels, for example, would be the way to go. However, I can see the desire for humans to create robots that resemble people. It has fascinated popular literature for much of the last century. I don't see that interest in the popular culture waning any time soon. It almost seems that the interface developed from this research would be the human resembling robot.
No comments:
Post a Comment