afterwards but don't end up asking

I still need time to go over the material

I can ask them

during but ...

I go over the material on my own time

I browse for alternative material

...I've tried everything and I'm still lost

...I'll deal with understanding later. Let's move on.

... we discuss them in lecture and/or tutorial

True

False (we must see the prediction at every time step)

False, RNNs are not classifiers

What's a time series?

is a learning algorithm

computes the EXACT gradients.

APPROXIMATES the gradients

is a classification model

is a learning algorithm

computes the exact gradients.

approximates the gradients

is a classification model

turning the RNN into a feedforward network

recomputing the EXACT same operation at t=1 over and over

Making an RNN go back in time.

The weight matrices between sucessive layers are duplicates of W

The W's between time steps are always different.

exponential forgetting

unfolding an RNN is very tricky

no one really understands how to implement an RNN

vanishing gradient problem

Both have to do with computational complexity.

(B): applies to feedforward networks only

Both lead to the same problem

(A): activity at t is insensitive to t=0 (B) multiplying small derivatives

W

V

U

x