[CS241] Asgn2 questions

[Lenora Huang]

Hi, everyone.
Can anyone help to explain what the question (d) mean?
For "all possible pairs", compute p(tense in sentence_i| "the" tense
appeared in some S_j j<i-1)
If it's for "all possible pairs", why specify "the" tense?

Also, when computing KL divergence, if the probability is 0, can we
ignore it because log(p) will be inf.?
And does the order of the probability matter? i.e, should we use
KL((b)||(c)) or KL((c)||(b))?

Thanks for answering in advance.

Lenora