Interactive Machine Learning
, Brigham Young University
Date: Friday, May 02, 2008
Time: 2:00 PM to 3:00 PM
Refreshments: 3:45 PM
Location: Patil/Kiva Seminar Room G449
Host: Rob Miller, MIT CSAIL
Contact: Michael Bernstein, x3-0452, email@example.com
Speaker URL: None
firstname.lastname@example.org, email@example.com, firstname.lastname@example.org
TALK: Interactive Machine Learning
Machine learning offers the promise of assisting users to create new tools simply by demonstrating the desired outcome. We have built one such tool "Image Processing with Crayons" for creating classifiers for image-based problems. The tool empowers a much larger class of people who can create image-based interactive techniques.
However, our observations of people using Crayons have pointed out several challenges in the way machine learning algorithms are designed. People do not behave in statistically uniform distributions and more importantly their interaction with the learning algorithm distorts their behavior in specific ways. The lessons we have learned will be discussed along new directions for machine learning algorithms that might learn faster in the face of user behavior.
Dan R. Olsen Jr. is a Professor of Computer Science at Brigham Young University. He was formerly the director of CMU's HCI Institute and founding editor of ACM's Transactions on Computer Human Interaction (TOCHI). For the last 25 years he has been working on software architectures and techniques to support the construction of user interfaces. His most recent work is in human-robot interaction and in architectures that integrate machine learning into the user interface.
Created by Linda L. Julien at Wednesday, June 19, 2013 at 6:23 AM.