Stanford CS224N: NLP with Deep Learning | Winter 2019 | Lecture 6 - Language Models and RNNs

[@kiran5918]

Handled the questions so well. Crisp and clear answers

[@kartiksirwani4657]

her lectures are on point and very clear.

[@adithyagiri7933]

Clear and crisp Lecture

[@saeedvahidian53]

Very clear and to-the-point lecture. Better than Chris!

[@paninilal8322]

Use of precise examples is very nice

[@stanfordonline]

Awesome feedback, thanks for watching!

[@danielsun4928]

seems better than Chris' talk

[@shrutiiyyer2783]

A question here! Will the RNN keep generating the same exact sequence ( during sequence generation and not training) if the starting word is the same. For example, if it randomly chooses 'the' to be the first word y0 will the rest of the sequence be the same if this generation was run twice given that both the times the first word generated was 'the'?

[@mohammedbouri8351]

Why the RNNs share the same weights at each time step? I didn't understand the goal behind it.

[@unknownhero6187]

What is the physical interpretation of the hidden state and corresponding weights matrix?

[@Teng_XD]

just linear combinations of vectors

[@paninilal8322]

It is the inner working of an intelligent unit like NN. Just like (actual) neurons inside the brain. A different subset(s) of neurons gets activated each time corresponding the respective input.