NOTE: This Tuesday we will meet in the CS Commons at noon. Please bring your own lunch; we'll try to have some drinks and snacks, too.
This week in CS Table, we will consider some applications of machine learning that show both potential benefit and great risks: The use of ML algorithms in decisions related to bail, sentencing, and parole. We have two readings:
On Thursday, September 6, John Stone will give a talk on adversarial examples, which are inputs to software applications for classification, assessment, or diagnosis that are specifically contrived to elicit incorrect or misleading results. Many applications based on neural networks configured by machine-learning algorithms have been found to be vulnerable to such examples. The talk will explain the nature of the vulnerability and explore possible explanations.
At 4:00 p.m., refreshments will be served in the Computer Science Commons. The talk, "Adversarial Examples; or, When Is a School Bus an Ostrich?" will begin at 4:15 p.m. in Noyce 3821. Everyone is welcome to attend!
On Thursday, October 29, Shitanshu Aggarwal 2011 and Jay Lidaka 2010 will present their summer research work, carred out under the direction of Professor Jerod Weinman.
Refreshments will be served at 4:15 p.m. in the Computer Science Commons
(Noyce 3817). The talk, Parallel training: speeding up machine learning
using graphical processing units,
will follow at 4:30 p.m. in Noyce
3821.
On Thursday, September 10, Jerod Weinman of the Department of Computer
Science will present the first talk in this year's Thursday
Extras
series, entitled Efficient machine learning for
computer vision-based depth perception.
Refreshments will be served at 4:15 p.m. in the Computer Science Commons (Noyce 3817). The talk will follow at 4:30 p.m. in Noyce 3821.