For part 2 in assignment 4

[prateekagrawal06]

Hi Professor,

In part 2 of the assignment.

Do you want us to train RNN on the data we get after vectorizing the make_feature_dict function?
Or do you want us to evaluate a totally different approach where we decide what kind of feature we want to go in RNN.

[aronwc]

No need to use make_feature_dict for the RNN. Either one-hot encoding or word2vec vectors as input to RNN.

…

On Apr 21, 2017, at 4:52 PM, prateekagrawal06 ***@***.***> wrote: Hi Professor, In part 2 of the assignment. Do you want us to train RNN on the data we get after vectorizing the make_feature_dict function? Or do you want us to evaluate a totally different approach where we decide what kind of feature we want to go in RNN. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.

[changediyasunny]

Hi Prof. @aronwc ,

could you please elaborate more with small example ? it is bit confusing as what should be the input and output ?

[aronwc]

Input are words, outputs are NER labels. Each word is a vector, either one hot, like our example in class, or the output of word2vec.

…

On Apr 22, 2017, at 1:52 AM, schanged ***@***.***> wrote: Hi Prof. @aronwc , could you please elaborate more with small example ? it is bit confusing as what should be the input and output ? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.