1
00:00:00,833 --> 00:00:01,166
Hello and

2
00:00:01,166 --> 00:00:04,400
welcome back to the course on Deep
Natural Language Processing.

3
00:00:04,400 --> 00:00:07,700
And today we're going to talk about
the types of natural language processing.

4
00:00:08,200 --> 00:00:10,733
So we've got two Venn diagrams here.

5
00:00:10,733 --> 00:00:13,500
Or we've got a Venn diagram
with two circles in it.

6
00:00:13,500 --> 00:00:17,033
And we are going to look

7
00:00:17,033 --> 00:00:21,033
at the different areas of,

8
00:00:21,633 --> 00:00:24,633
natural language processing
that are going to come up in this course.

9
00:00:25,000 --> 00:00:29,200
So on the left we've got natural language
processing overall.

10
00:00:29,566 --> 00:00:33,200
And this refers
to the whole circle on the left.

11
00:00:33,566 --> 00:00:37,100
So the reason why we've called it in
just this green part is because that's the

12
00:00:37,133 --> 00:00:38,500
non-overlapping part.

13
00:00:38,500 --> 00:00:43,033
So we know that anything in here,
is just natural language processing.

14
00:00:43,033 --> 00:00:46,100
We followed with disregard to this
second circle.

15
00:00:46,233 --> 00:00:50,700
But, natural language processing is indeed
everything that is in this first circle.

16
00:00:51,366 --> 00:00:54,633
Then we've got on the right deep learning.

17
00:00:54,633 --> 00:00:58,300
So these are all algorithms
that have something

18
00:00:58,300 --> 00:01:01,400
to do with neural networks, deep learning.

19
00:01:01,800 --> 00:01:04,800
basically anything that's called a deep
learning algorithm falls in here.

20
00:01:04,800 --> 00:01:07,900
They don't have to be natural
language processing.

21
00:01:07,900 --> 00:01:10,566
They can be, classification.

22
00:01:10,566 --> 00:01:11,633
they can be anything.

23
00:01:11,633 --> 00:01:14,033
So they can be
that's in deep learning or here.

24
00:01:14,033 --> 00:01:17,500
And natural language
processing is any algorithm, any model

25
00:01:17,500 --> 00:01:20,666
that has something to do
with processing of natural language,

26
00:01:21,233 --> 00:01:24,000
into machine terms.

27
00:01:25,233 --> 00:01:26,000
And then finally

28
00:01:26,000 --> 00:01:29,000
in the overlap we have deep NLP.

29
00:01:29,133 --> 00:01:33,200
So these are models which have to do
with natural language processing,

30
00:01:33,366 --> 00:01:37,600
but also which are deep learning models
which are neural network models.

31
00:01:38,100 --> 00:01:39,300
And yeah.

32
00:01:39,300 --> 00:01:42,300
So that's the part
that we're going to be aiming for.

33
00:01:42,333 --> 00:01:46,866
But it's also very good
to have a visibility of all three,

34
00:01:47,300 --> 00:01:52,400
because in this course we will be talking
about some models that fall just in here.

35
00:01:52,400 --> 00:01:54,233
And then we'll be talking about models
here.

36
00:01:54,233 --> 00:01:57,900
And it'll be good to compare and see
how the world has changed over time,

37
00:01:57,900 --> 00:02:01,400
and why these models are often
better than these models.

38
00:02:02,600 --> 00:02:06,200
and the other thing to note here
is that the size of these diagrams

39
00:02:06,500 --> 00:02:12,900
is not reflective of the importance
or the volumes of these different fields.

40
00:02:12,900 --> 00:02:17,533
So, I just, I set circles of the same size
simply because we want

41
00:02:17,533 --> 00:02:21,300
a visual representation of, of the overlap
and that these fields exist.

42
00:02:21,666 --> 00:02:24,066
But don't, take the size into account.

43
00:02:24,066 --> 00:02:26,166
It's not to scale at all.

44
00:02:27,133 --> 00:02:31,666
And, finally, there is another part,

45
00:02:31,966 --> 00:02:36,300
another part of this event diagram,
which is very important to us.

46
00:02:36,700 --> 00:02:39,700
And it is this part over here, a

47
00:02:39,833 --> 00:02:45,300
sub, section of a deep
NLP called a sequence to sequence.

48
00:02:45,300 --> 00:02:48,533
So sequence to sequence
models are the most cutting edge, the most

49
00:02:48,533 --> 00:02:52,833
powerful models that exist right now
for, natural language processing.

50
00:02:52,833 --> 00:02:55,833
And that's
what we're going to be looking at.

51
00:02:56,200 --> 00:02:58,800
So as you'll see throughout this course,
we will make our way

52
00:02:58,800 --> 00:03:02,833
through the natural language
processing side of things into deep NLP.

53
00:03:02,833 --> 00:03:05,566
And then we will go into sequence
to sequence.

54
00:03:05,566 --> 00:03:07,766
It'll be a fun and exciting journey.

55
00:03:07,766 --> 00:03:11,200
and the other thing that I wanted to
mention is you will also notice

56
00:03:11,200 --> 00:03:14,266
that throughout this course,
even though it's focused

57
00:03:14,266 --> 00:03:17,400
on chatbots,
we won't be talking about just chat bots.

58
00:03:17,700 --> 00:03:22,400
We'll be looking at different examples of,
how these models

59
00:03:22,400 --> 00:03:26,200
from here and from here and from here
can be applied to different things.

60
00:03:26,200 --> 00:03:29,100
Because the applications are huge.

61
00:03:29,100 --> 00:03:34,000
We can talk, we can apply them in,
natural, in neural machine translation.

62
00:03:34,000 --> 00:03:37,100
We can apply them, in image captioning.

63
00:03:37,100 --> 00:03:40,100
We can apply them in, speech recognition,

64
00:03:40,233 --> 00:03:43,933
questions and answers, text
summarization, lots and lots of models.

65
00:03:43,933 --> 00:03:49,233
So, we will be looking at different ones
and they will be of different types.

66
00:03:49,500 --> 00:03:52,733
So this map will come in handy
as we go through the course.

67
00:03:52,733 --> 00:03:54,500
And it will be popping up here and there.

68
00:03:54,500 --> 00:04:00,100
So I think it was very important
for us to set the foundation, right.

69
00:04:00,100 --> 00:04:01,966
So that now we're ready to proceed.

70
00:04:01,966 --> 00:04:04,166
And I can't wait to see you
on the next tutorial.

71
00:04:04,166 --> 00:04:07,666
And until then, enjoy a deep
and natural language processing.