1
00:00:00,000 --> 00:00:08,090
Now let's see some applications that
actually use big data and AI techniques at

2
00:00:08,090 --> 00:00:16,029
web scale to achieve the prevalent of
predictive intelligence that we, human

3
00:00:16,029 --> 00:00:24,285
exhibits in our every day lives.
The first and most prevalent example is

4
00:00:24,285 --> 00:00:32,663
online advertising, where predicting our
intent when we search or read documents on

5
00:00:32,663 --> 00:00:39,773
the web and our interests is a part of
every day life today.

6
00:00:39,773 --> 00:00:48,626
Similarly gauging consumer sentiment from
the multitude of conversations which are

7
00:00:48,626 --> 00:00:55,353
not public on Twitter and even some
Facebook forums and based on that,

8
00:00:55,353 --> 00:01:02,570
predicting our own behavior is something
many organizations are already doing

9
00:01:02,570 --> 00:01:11,754
without our actually being aware.
On the other side, adverse events such as

10
00:01:11,754 --> 00:01:20,558
fires, strikes, floods, and even larger
disasters such as earthquakes, reach

11
00:01:20,558 --> 00:01:26,792
Twitter first before hitting any news
channel or print media.

12
00:01:26,792 --> 00:01:35,035
And organizations are now beginning to tap
into detecting such events as soon as

13
00:01:35,035 --> 00:01:43,833
possible and rapidly predicting their
impact, so that they can react even

14
00:01:43,833 --> 00:01:50,135
faster.
Intelligent question answering, such as in

15
00:01:50,135 --> 00:01:57,435
the Watson Program that beat the Jeopardy
champions, is another example of

16
00:01:57,435 --> 00:02:06,431
intelligence coming from, processing large
volumes of data garnered from the web.

17
00:02:06,431 --> 00:02:14,832
Machines are also now able to categorize
and recognize places, faces and people

18
00:02:14,832 --> 00:02:23,642
just as we humans do so every day because
there are a large volume of images and

19
00:02:23,642 --> 00:02:29,556
videos available to process and learn
from.

20
00:02:29,556 --> 00:02:38,871
In the future, personalize gnomic medicine
might become a reality and it's already

21
00:02:38,871 --> 00:02:48,260
beginning to be explored as more and more
people share their DNA samples to get some

22
00:02:48,260 --> 00:02:58,641
understanding of their ancestry or their
potential probability of contracting some

23
00:02:58,641 --> 00:03:05,145
genetic diseases.
As more and more such data gets shared,

24
00:03:05,145 --> 00:03:12,526
when it's quarterly date with clinical
data about the actual effectiveness of

25
00:03:12,526 --> 00:03:19,785
different medicines on different kinds of
genetic profiles, we can potentially see

26
00:03:19,785 --> 00:03:25,498
vast volumes of data leading to better
medication for all of us.

27
00:03:25,498 --> 00:03:35,024
Similar examples are possible in other
arenas such as more intelligent

28
00:03:35,024 --> 00:03:43,099
distribution and consumption of energy,
water and other scarce resources using

29
00:03:43,099 --> 00:03:53,008
intelligent sensors and deep analytics on
the large volumes of data that one can

30
00:03:53,008 --> 00:03:59,548
collect.
Securing ourselves better from bad guys is

31
00:03:59,548 --> 00:04:07,883
another application of web intelligence
that is happening all the time, largely

32
00:04:07,883 --> 00:04:14,278
without our knowing it.
Another term attracting a lot of attention

33
00:04:14,278 --> 00:04:21,446
these days is big data analytics.
To a certain extent, this is all about

34
00:04:21,446 --> 00:04:26,575
large enterprises.
Which are outside the web world

35
00:04:26,575 --> 00:04:33,926
traditionally, trying to exploit more
efficient technology which was developed

36
00:04:33,926 --> 00:04:40,018
by the web companies for their large scale
web intelligence tasks.

37
00:04:40,018 --> 00:04:47,051
Technology which is proving to be equally
good if not better and cheaper than

38
00:04:47,051 --> 00:04:52,825
traditional database technology.
Equal importantly though.

39
00:04:52,825 --> 00:05:00,416
Big data analytics is about fusing the
social intelligence available from

40
00:05:00,416 --> 00:05:07,794
external sources with the business
intelligence available from internal data

41
00:05:07,794 --> 00:05:12,901
sources available within every large
enterprise.

42
00:05:12,901 --> 00:05:22,351
In other words, a mix of private data and
web data on which the web intelligence

43
00:05:22,351 --> 00:05:30,258
techniques used in the kinds of
applications we spoke about a few minutes

44
00:05:30,258 --> 00:05:35,549
ago.
As a result, better sales and marketing,

45
00:05:35,549 --> 00:05:44,705
more intelligence supply chains, and in
general digitally enabled, mobile enabled,

46
00:05:44,705 --> 00:05:51,572
data driven business models and processes
are becoming possible using the same

47
00:05:51,572 --> 00:05:58,251
techniques used by the web companies to
understand all of us better.

48
00:05:58,251 --> 00:06:06,224
So in a nutshell, bit data analytics is
all about brick and mortar firms trying to

49
00:06:06,224 --> 00:06:13,006
emulate the web companies by using exactly
the same techniques that they use for web