1
00:00:00,350 --> 00:00:02,410
Next, we have Amazon Personalize,

2
00:00:02,410 --> 00:00:05,330
which is a fully machine learning service to build apps

3
00:00:05,330 --> 00:00:08,100
with real-time personalized recommendations.

4
00:00:08,100 --> 00:00:09,260
So what could be a recommendation?

5
00:00:09,260 --> 00:00:11,880
For example, a personalized product recommendation,

6
00:00:11,880 --> 00:00:14,320
or re-ranking, or customized direct marketing.

7
00:00:14,320 --> 00:00:17,410
For example, a user has bought a lot of gardening tools,

8
00:00:17,410 --> 00:00:19,470
and you want to provide recommendations

9
00:00:19,470 --> 00:00:23,700
on the next one to buy based on a personalized service.

10
00:00:23,700 --> 00:00:25,820
So this is the same technology used by Amazon.com.

11
00:00:25,820 --> 00:00:28,160
So when you go and shop on Amazon.com

12
00:00:28,160 --> 00:00:30,330
and after buying a few products, what you will see

13
00:00:30,330 --> 00:00:32,729
is that Amazon.com start recommending products

14
00:00:32,729 --> 00:00:35,890
in the same category or in completely different categories

15
00:00:35,890 --> 00:00:38,300
based on how you've been searching

16
00:00:38,300 --> 00:00:39,740
and how you've been buying

17
00:00:39,740 --> 00:00:42,190
and user interest and that kind of things.

18
00:00:42,190 --> 00:00:44,720
So Personalize is how you access this from within AWS.

19
00:00:44,720 --> 00:00:48,100
So, you read your input data from Amazon S3.

20
00:00:48,100 --> 00:00:50,110
For example, it could be user interactions,

21
00:00:50,110 --> 00:00:51,470
those kind of things.

22
00:00:51,470 --> 00:00:54,550
Also, you can use the Amazon Personalize API

23
00:00:54,550 --> 00:00:56,460
to have real-time data integration

24
00:00:56,460 --> 00:00:59,040
into the Amazon Personalize service.

25
00:00:59,040 --> 00:01:02,920
And then this will expose a customized personalized API

26
00:01:02,920 --> 00:01:04,650
for your websites and applications,

27
00:01:04,650 --> 00:01:06,030
your mobile applications.

28
00:01:06,030 --> 00:01:08,710
Also, you can send SMS or emails

29
00:01:08,710 --> 00:01:10,810
for personalization as well.

30
00:01:10,810 --> 00:01:12,480
So you have all these integrations.

31
00:01:12,480 --> 00:01:15,200
It takes days, not months, to build this model.

32
00:01:15,200 --> 00:01:18,090
So you don't need to build, train, and deploy ML solutions.

33
00:01:18,090 --> 00:01:20,680
You can just use this bundled as is.

34
00:01:20,680 --> 00:01:23,320
And so the use cases is going to be retail stores,

35
00:01:23,320 --> 00:01:25,060
and media, and entertainment.

36
00:01:25,060 --> 00:01:26,250
So from an exam perspective,

37
00:01:26,250 --> 00:01:28,440
anytime you see a machine learning service

38
00:01:28,440 --> 00:01:31,760
to build recommendations and personalized recommendations,

39
00:01:31,760 --> 00:01:34,040
think Amazon Personalize, that's it.

40
00:01:34,040 --> 00:01:35,790
I will see you in the next lecture.