1
00:00:00,000 --> 00:00:05,051
Welcome to this course on Web Intelligence
and Big Data.

2
00:00:05,086 --> 00:00:13,212
In this first lecture, I'll introduce you
to the topic, as well as outlined the,

3
00:00:13,212 --> 00:00:18,848
course itself and discuss some of the
logistics and evaluation.

4
00:00:18,848 --> 00:00:25,840
When talking about intelligence, one can't
help but recall Allen Turing and his

5
00:00:25,840 --> 00:00:33,385
famous thought experiment introduced way
be back in 1950, to test if a computer was

6
00:00:33,385 --> 00:00:38,376
actually intelligent.
Turing's original test is based on a

7
00:00:38,376 --> 00:00:45,586
popular party game where a human judge is
able to converse using typewritten text

8
00:00:45,586 --> 00:00:50,066
only between two people, one a man and one
a woman.

9
00:00:50,066 --> 00:00:56,082
And needs to judge, based on the
conversation alone, which is a man and

10
00:00:56,082 --> 00:01:02,068
which is a woman.
Turing's variation of this imitation game

11
00:01:02,068 --> 00:01:07,002
was to replace one of the participants by
a machine.

12
00:01:07,002 --> 00:01:13,504
And now the judge needs to determine based
on the conversations alone, who is human

13
00:01:13,504 --> 00:01:18,096
and who is a machine.
Turing claimed that if a machine could

14
00:01:18,096 --> 00:01:25,039
successfully fool a human judge into
believing that it was not a machine and

15
00:01:25,039 --> 00:01:31,040
actually human, then that machine should
be considered to be intelligent.

16
00:01:32,056 --> 00:01:38,416
An interesting variation of the Turing
test is where the judge's, now a machine,

17
00:01:38,416 --> 00:01:44,395
and needs to figure out which of the
parties it's talking to is a human, and

18
00:01:44,395 --> 00:01:49,614
which is another machine.
Now, do you think you might have seen or

19
00:01:49,614 --> 00:01:53,093
experienced such a reverse Turing test
yourselves?

20
00:01:54,095 --> 00:02:01,001
Think about the captchas which all of us
encounter on various webpages.

21
00:02:01,001 --> 00:02:07,082
They ask us to type in a set of letters
looking at disfigured versions of those

22
00:02:07,082 --> 00:02:11,015
words.
A task which machines find quite

23
00:02:11,015 --> 00:02:15,066
difficult.
So this is an example where the machine is

24
00:02:15,066 --> 00:02:20,001
trying to figure out whether we are
actually human.

25
00:02:20,001 --> 00:02:24,087
And so it's the reverse hearing test in
action every day.

26
00:02:24,087 --> 00:02:30,000
But there's more.
As we shall soon see in this course,

27
00:02:30,000 --> 00:02:37,098
machines inside the web can pretty easily
figure out based on the conversations that

28
00:02:37,098 --> 00:02:46,145
we have with each other whether we are
male or female, the products we like and

29
00:02:46,145 --> 00:02:54,218
dislike, whether we are shopping or just
surfing, whether we are rich or poor, or

30
00:02:54,218 --> 00:03:01,928
even scary things like our ethnicity.
The surprising thing is, that this is

31
00:03:01,928 --> 00:03:06,666
happening all the time on the web, as we
shall see in our course.

32
00:03:06,666 --> 00:03:09,063
So is the web becoming intelligent.