NLP course lesson 3

Deep Learning
1

I’m learning the NLP course, and I saw in lesson 3 that Ngrams and Naive Bayes is the old-fashioned way.
Should I dig deeper to understand the algorithms clearly or just know the basics and going to the next lesson?
Thanks in advance.

2

Hi Hao
I think it is best to learn just what is on the course and then later learn about the other things.
For Bayes
The one thing I had to hunt down is Gamma(n+1) = n * (n-1) * (n-2)… so n!
Beta(a,b) = Gamma(a)*Gamma(b)/Gamma(a+b). This is used when you have no prior data.

Here are some articles

Regards Conwyn

3

you mean if the lesson told like SVD or Naive Bayes, I should search for more information about it and go to the next lesson ?

4

Hi Hao
No the opposite. Complete the course and then go and read about SVD and Bayes.
Sorry for my lack of clarity.
Regards Conwyn

5

Thanks a lot for your responses, so I’m ready for the next lesson now.

6

Hi Hao
It is worth doing the Computational Linear Algebra after NLP. It gives the background to SVD. It is mathematical and the audio is poor but I think if you do every single Fast.AI courses it helps with the overall understanding.
Regards Conwyn

7

Thanks for your advice. I really appreciate it.