Please answer both questions completely and throughly with citations wherever required.
1.
We want to build a naive Bayes sentiment classifier using add-1 smoothing. Here is our training set:
– the movie has no plot
– honestly pretty boring
+ pretty interesting movie
Test Set:
pretty enjoyable plot
Answer the questions below given the sets above:
1.
Compute whether the sentence in the test set is of class positive or negative (you may
need a computer for this final computation).
2.
Read the following article:
Wu, Yonghui, et al. (2016). Google’s Neural Machine Translation System: Bridging the Gap between Human
and Machine Translation.
Then, describe the model architecture presented in figure 1 with some real numbers:
1. left encoder network
1.
a. what are the GPUs layers about: explain inputs/outputs
b. what is Encoder LSTM layer about
2. middle attention module
3. right decoder network
a.
a. explain Softmax with example
b. what is Decoder LSTM layer about
You will need the following articles to help you with descriptions:
• Bahdanau, Dzmitry & Cho, Kyunghyun & Bengio, Y. (2014). Neural Machine Translation by
Jointly Learning to Align and Translate. ArXiv. 1409.
• Sutskever, Ilya & Vinyals, Oriol & Le, Quoc. (2014). Sequence to Sequence Learning with Neural
Networks. Advances in Neural Information Processing Systems. 4.