CS 451 class prep 30

1. Work on your final project. Recall that progress reports are due on Wed. As a rough guideline, I expect that each of you puts in 4-8 hours of work on your final project by Wed.

2. Watch the following Coursera Deep Learning 2 videos in preparation for the next class.

Key: * = required
     ? = optional

Course 2: Improving Deep Neural Networks:
Hyperparameter tuning, Regularization and Optimization

WEEK 1
Practical aspects of Deep Learning

?  12 Video: Train / Dev / Test sets    [you do need to know what "dev set" means]
?   8 Video: Bias / Variance
*   6 Video: Basic Recipe for Machine Learning
?   9 Video: Regularization
?   7 Video: Why regularization reduces overfitting?
*   9 Video: Dropout Regularization
*   7 Video: Understanding Dropout
*   8 Video: Other regularization methods
?   5 Video: Normalizing inputs
*   6 Video: Vanishing / Exploding gradients
?   6 Video: Weight Initialization for Deep Networks
?   6 Video: Numerical approximation of gradients
?   6 Video: Gradient checking
?   5 Video: Gradient Checking Implementation Notes

Heroes of Deep Learning
?  25 Video: Yoshua Bengio interview

*  36    min
?  64+25 min