Monday
April 28
4:00 - 4:50 PM 
Kelley 1001

 

Weng-Keen Wong 
Assistant Professor
School of Electrical Engineering & Computer Science
Oregon State University

End-user debugging of machine learning systems 

How do you debug a program that was written by a machine instead of a
person, especially when you do not know much about programming or
machine learning, and are working with a program you cannot even see?
This is the problem faced by users of a new type of program being used
today, namely, machine learning systems that, after being deployed,
customize themselves by learning from an end-user's behavior. Prime
examples of these programs include adaptive user interfaces, intelligent
desktop assistants, email classifiers, and recommender systems.
Inaccurate predictions by these learning systems erode users' trust and
curb widespread acceptance of such systems. We seek to improve both the
performance and acceptance of machine learning systems by allowing
end-users to debug these learned programs when they make incorrect
predictions. Achieving this goal requires two key components: (1)
providing explanations of the system's reasoning that is suitable for
end-users untrained in machine learning and (2) improving the
performance of the learned program by incorporating corrective feedback
from the end-user. This talk will present our work on these two
components along with results from a user study in which participants
were allowed to provide corrective feedback to an email classifier. 

Biography

Weng-Keen Wong is an Assistant Professor of Computer Science at Oregon
State University. Prior to being a faculty member at OSU, he was a
post-doctoral associate at the Center for Biomedical Informatics at the
University of Pittsburgh. He received his Ph.D. (2004) and M.S. (2001)
in Computer Science from Carnegie Mellon University, and his B.Sc.
(1997) from the University of British Columbia. His Ph.D. thesis was
entitled "Data Mining Algorithms for the Early Detection of Disease
Outbreaks". Currently, his research areas are in data mining and machine
learning, with specific interests in anomaly detection,
human-in-the-loop learning, and applications of machine learning to
ecological problems. 

 

_______________________________________________
Colloquium mailing list
[email protected]
https://secure.engr.oregonstate.edu/mailman/listinfo/colloquium

Reply via email to