1
00:00:00,126 --> 00:00:03,971
In the earlier courses in this
specialization, you got an introduction to

2
00:00:03,971 --> 00:00:08,319
machine learning and deep learning with
a heavy focus on computer vision problems.

3
00:00:08,319 --> 00:00:10,361
You learned about neural networks and

4
00:00:10,361 --> 00:00:13,502
how they can match patterns
to perform classifications.

5
00:00:13,502 --> 00:00:15,858
And then how you can
give them new data and

6
00:00:15,858 --> 00:00:18,500
have them predict what
they might be seeing.

7
00:00:18,500 --> 00:00:21,567
You learned how to make
that a little smarter for

8
00:00:21,567 --> 00:00:26,016
images using convolutions to identify
the features in the images and

9
00:00:26,016 --> 00:00:30,178
classify based on those instead
of just matching on raw pixels.

10
00:00:30,178 --> 00:00:32,073
This helped you classify images for

11
00:00:32,073 --> 00:00:36,124
more real world style ones instead of
using a very controlled environment.

12
00:00:37,430 --> 00:00:40,375
In this course we're going to
go back to building models but

13
00:00:40,375 --> 00:00:45,035
we'll focused on text and how you can
build classifier is based on text models.

14
00:00:45,035 --> 00:00:48,410
We'll start by looking at
sentiment in text, and

15
00:00:48,410 --> 00:00:53,180
learn how to build models that understand
text that are trained on labeled text, and

16
00:00:53,180 --> 00:00:56,250
then can then classify new text
based on what they've seen.

17
00:00:57,360 --> 00:01:00,598
When we were dealing with images,
it was relatively easy for

18
00:01:00,598 --> 00:01:04,925
us to feed them into a neural network,
as the pixel values were already numbers.

19
00:01:04,925 --> 00:01:09,405
And the network could learn parameters
of functions that could be used to fit

20
00:01:09,405 --> 00:01:10,599
classes to labels.

21
00:01:10,599 --> 00:01:12,231
But what happens with text?

22
00:01:12,231 --> 00:01:14,760
How can we do that with sentences and
words?