leetcode word break# JobHunting - 待字闺中
s*w
1 楼
much easier than previous paired up I, II
I encountered it twice in my recent interviews and they will follow up with
questions if you have machine learning relevant background
1. how to pick the best out of many ways of decompositions?
unigram, bi-gram, tri-gram, etc.
2. how exactly do you train such models?
essentially the same as q1, just to make sure you know more than the phrase
itself
3. what happens if some word in testing set does not appear in training set?
how would you correct it? do you have any math to back it up?
this smoothing question is asked every time if the problem setting is close
I encountered it twice in my recent interviews and they will follow up with
questions if you have machine learning relevant background
1. how to pick the best out of many ways of decompositions?
unigram, bi-gram, tri-gram, etc.
2. how exactly do you train such models?
essentially the same as q1, just to make sure you know more than the phrase
itself
3. what happens if some word in testing set does not appear in training set?
how would you correct it? do you have any math to back it up?
this smoothing question is asked every time if the problem setting is close