02.23 Lec 6

Post date: Feb 22, 2017 5:29:31 AM

Lecture: Underfitting and going beyond linear models. Overfitting, regularization and dropout.

Reading: Multi-layer perceptrons, Dropout, Regularization, IJulia notebook

Lab: Please try to sign up at AWS within the week and email comp541group@ku.edu.tr if you have any problems. It may take a couple of days, and we will use AWS in the lab on Friday. You can refer to the CIT documentation and the Knet documentation for instructions. I recommend using emacs for development on text based terminal interfaces. Try "Ctrl-h" followed by "t" in emacs to launch a tutorial.

Project: Reading your paper, first pass due Feb 26 on Classroom.