en_Intro to Machine Learning.srt

0
00:00:00,680 --> 00:00:02,530
Hello, and welcome!

1
00:00:02,530 --> 00:00:06,839
In this video I will give you a high level introduction to Machine Learning.

2
00:00:06,839 --> 00:00:09,480
So let’s get started.

3
00:00:09,480 --> 00:00:13,520
This is a human cell sample extracted from a patient.

4
00:00:13,520 --> 00:00:20,240
And this cell has characteristics … for example, its Clump thickness is 6, its Uniformity

5
00:00:20,240 --> 00:00:26,660
of cell size is 1, its Marginal adhesion is 1, and so on.

6
00:00:26,660 --> 00:00:32,910
One of the interesting questions we can ask, at this point is: "Is this a Benign or Malignant

7
00:00:32,910 --> 00:00:34,270
cell?"

8
00:00:34,270 --> 00:00:39,750
In contrast with a benign tumor, a malignant tumor is a tumor that may invade its surrounding

9
00:00:39,750 --> 00:00:46,660
tissue or spread around the body, and diagnosing it early might be the key to a patient’s

10
00:00:46,660 --> 00:00:47,660
survival.

11
00:00:47,660 --> 00:00:53,000
One could easily presume that only a doctor with years of experience could diagnose that

12
00:00:53,000 --> 00:00:57,510
tumor and say if the patient is developing cancer or not.

13
00:00:57,510 --> 00:00:58,670
Right?

14
00:00:58,670 --> 00:01:04,360
Well, imagine that you’ve obtained a dataset containing characteristics of thousands of

15
00:01:04,360 --> 00:01:09,720
human cell samples extracted from patients who were believed to be at risk of developing

16
00:01:09,720 --> 00:01:11,530
cancer.

17
00:01:11,530 --> 00:01:16,850
Analysis of the original data showed that many of the characteristics differed significantly

18
00:01:16,850 --> 00:01:20,430
between benign and malignant samples.

19
00:01:20,430 --> 00:01:26,290
You can use the values of these cell characteristics in samples from other patients to give an

20
00:01:26,290 --> 00:01:31,200
early indication of whether a new sample might be benign or malignant.

21
00:01:31,200 --> 00:01:37,540
You should clean your data, select a proper algorithm for building a prediction model,

22
00:01:37,540 --> 00:01:43,799
and train your model to understand patterns of benign or malignant cells within the data.

23
00:01:43,799 --> 00:01:48,850
Once the model has been trained by going through data iteratively, it can be used to predict

24
00:01:48,850 --> 00:01:53,880
your new or unknown cell with a rather high accuracy.

25
00:01:53,880 --> 00:01:56,140
This is machine learning!

26
00:01:56,140 --> 00:02:00,920
It is the way that a machine learning model can do a doctor’s task or at least help

27
00:02:00,920 --> 00:02:04,060
that doctor make the process faster.

28
00:02:04,060 --> 00:02:09,280
Now, let me give a formal definition of machine learning.

29
00:02:09,280 --> 00:02:13,909
Machine learning is the subfield of computer science that gives "computers the ability

30
00:02:13,909 --> 00:02:18,290
to learn without being explicitly programmed.”

31
00:02:18,290 --> 00:02:24,159
Let me explain what I mean when I say “without being explicitly programmed.”

32
00:02:24,159 --> 00:02:29,500
Assume that you have a dataset of images of animals such as cats and dogs, and you want

33
00:02:29,500 --> 00:02:35,529
to have software or an application that can recognize and differentiate them.

34
00:02:35,529 --> 00:02:41,739
The first thing that you have to do here is interpret the images as a set of feature sets.

35
00:02:41,739 --> 00:02:45,900
For example, does the image show the animal’s eyes?

36
00:02:45,900 --> 00:02:48,379
If so, what is their size?

37
00:02:48,379 --> 00:02:49,889
Does it have ears?

38
00:02:49,889 --> 00:02:51,370
What about a tail?

39
00:02:51,370 --> 00:02:52,370
How many legs?

40
00:02:52,370 --> 00:02:55,120
Does it have wings?

41
00:02:55,120 --> 00:03:00,400
Prior to machine learning, each image would be transformed to a vector of features.

42
00:03:00,400 --> 00:03:05,760
Then, traditionally, we had to write down some rules or methods in order to get computers

43
00:03:05,760 --> 00:03:08,889
to be intelligent and detect the animals.

44
00:03:08,889 --> 00:03:11,339
But, it was a failure.

45
00:03:11,339 --> 00:03:12,339
Why?

46
00:03:12,339 --> 00:03:18,879
Well, as you can guess, it needed a lot of rules, highly dependent on the current dataset,

47
00:03:18,879 --> 00:03:23,489
and not generalized enough to detect out-of-sample cases.

48
00:03:23,489 --> 00:03:27,040
This is when machine learning entered the scene.

49
00:03:27,040 --> 00:03:32,370
Using machine learning allows us to build a model that looks at all the feature sets,

50
00:03:32,370 --> 00:03:39,239
and their corresponding type of animals, and learn it learns the pattern of each animal.

51
00:03:39,239 --> 00:03:43,159
It is a model built by machine learning algorithms.

52
00:03:43,159 --> 00:03:47,049
It detects without explicitly being programmed to do so.

53
00:03:47,049 --> 00:03:53,760
In essence, machine learning follows the same process that a 4-year-old child uses to learn,

54
00:03:53,760 --> 00:03:56,859
understand, and differentiate animals.

55
00:03:56,859 --> 00:04:03,430
So, machine learning algorithms, inspired by the human learning process, iteratively

56
00:04:03,430 --> 00:04:09,159
learn from data, and allow computers to find hidden insights.

57
00:04:09,159 --> 00:04:15,530
These models help us in a variety of tasks, such as object recognition, summarization,

58
00:04:15,530 --> 00:04:18,989
recommendation, and so on.

59
00:04:18,988 --> 00:04:22,560
Machine Learning impacts society in a very influential way.

60
00:04:22,560 --> 00:04:26,919
Here are some real-life examples.

61
00:04:26,919 --> 00:04:33,410
First, how do you think Netflix and Amazon recommend videos, movies, and TV shows to its users?

62
00:04:33,410 --> 00:04:38,139
They use Machine Learning to produce suggestions that you might enjoy!

63
00:04:38,139 --> 00:04:42,439
This is similar to how your friends might recommend a television show to you, based

64
00:04:42,439 --> 00:04:45,550
on their knowledge of the types of shows you like to watch.

65
00:04:45,550 --> 00:04:50,880
How do you think banks make a decision when approving a loan application?

66
00:04:50,880 --> 00:04:57,169
They use machine learning to predict the probability of default for each applicant, and then approve

67
00:04:57,169 --> 00:05:01,700
or refuse the loan application based on that probability.

68
00:05:01,700 --> 00:05:08,199
Telecommunication companies use their customers’ demographic data to segment them, or predict

69
00:05:08,199 --> 00:05:12,540
if they will unsubscribe from their company the next month.

70
00:05:12,540 --> 00:05:17,330
There are many other applications of machine learning that we see every day in our daily

71
00:05:17,330 --> 00:05:25,500
life, such as chatbots, logging into our phones or even computer games using face recognition.

72
00:05:25,500 --> 00:05:29,229
Each of these use different machine learning techniques and algorithms.

73
00:05:29,229 --> 00:05:33,969
So, let’s quickly examine a few of the more popular techniques.

74
00:05:33,969 --> 00:05:40,080
The Regression/Estimation technique is used for predicting a continuous value, for example,

75
00:05:40,080 --> 00:05:45,389
predicting things like the price of a house based on its characteristics, or to estimate

76
00:05:45,389 --> 00:05:48,870
the Co2 emission from a car’s engine.

77
00:05:48,870 --> 00:05:54,580
A Classification technique is used for Predicting the class or category of a case, for example,

78
00:05:54,580 --> 00:06:01,030
if a cell is benign or malignant, or whether or not a customer will churn.

79
00:06:01,030 --> 00:06:06,910
Clustering groups of similar cases, for example, can find similar patients, or can be used

80
00:06:06,910 --> 00:06:10,830
for customer segmentation in the banking field.

81
00:06:10,830 --> 00:06:16,469
Association technique is used for finding items or events that often co-occur, for example,

82
00:06:16,469 --> 00:06:21,090
grocery items that are usually bought together by a particular customer.

83
00:06:21,090 --> 00:06:27,249
Anomaly detection is used to discover abnormal and unusual cases, for example, it is used

84
00:06:27,249 --> 00:06:30,150
for credit card fraud detection.

85
00:06:30,150 --> 00:06:35,650
Sequence mining is used for predicting the next event, for instance, the click-stream

86
00:06:35,650 --> 00:06:37,439
in websites.

87
00:06:37,439 --> 00:06:41,710
Dimension reduction is used to reduce the size of data.

88
00:06:41,710 --> 00:06:47,509
And finally, recommendation systems; this associates people's preferences with others

89
00:06:47,509 --> 00:06:54,289
who have similar tastes, and recommends new items to them, such as books or movies.

90
00:06:54,289 --> 00:06:58,530
We will cover some of these techniques in the next videos.

91
00:06:58,530 --> 00:07:03,789
By this point, I’m quite sure this question has crossed your mind, “What is the difference

92
00:07:03,789 --> 00:07:09,330
between these buzzwords that we keep hearing these days, such as Artificial intelligence

93
00:07:09,330 --> 00:07:13,040
(or AI), Machine Learning and Deep Learning?”

94
00:07:13,040 --> 00:07:16,520
Well, let me explain what is different between them.

95
00:07:16,520 --> 00:07:23,150
In brief, AI tries to make computers intelligent in order to mimic the cognitive functions

96
00:07:23,150 --> 00:07:24,639
of humans.

97
00:07:24,639 --> 00:07:31,440
So, Artificial Intelligence is a general field with a broad scope including: Computer Vision,

98
00:07:31,440 --> 00:07:36,830
Language Processing, Creativity, and Summarization.

99
00:07:36,830 --> 00:07:42,469
Machine Learning is the branch of AI that covers the statistical part of artificial

100
00:07:42,469 --> 00:07:43,469
intelligence.

101
00:07:43,469 --> 00:07:49,099
It teaches the computer to solve problems by looking at hundreds or thousands of examples,

102
00:07:49,099 --> 00:07:55,960
learning from them, and then using that experience to solve the same problem in new situations.

103
00:07:55,960 --> 00:08:01,219
And Deep Learning is a very special field of Machine Learning where computers can actually

104
00:08:01,219 --> 00:08:05,999
learn and make intelligent decisions on their own.

105
00:08:05,999 --> 00:08:12,919
Deep learning involves a deeper level of automation in comparison with most machine learning algorithms.

106
00:08:12,919 --> 00:08:17,550
Now that we’ve completed the introduction to Machine Learning, subsequent videos will

107
00:08:17,550 --> 00:08:23,180
focus on reviewing two main components: First, you’ll be learning about the purpose

108
00:08:23,180 --> 00:08:27,680
of Machine Learning and where it can be applied in the real world; and

109
00:08:27,680 --> 00:08:32,990
Second, you’ll get a general overview of Machine Learning topics, such as supervised

110
00:08:32,990 --> 00:08:39,370
vs unsupervised learning, model evaluation and various Machine Learning algorithms.

111
00:08:39,370 --> 00:08:44,481
So now that you have a sense with what’s in store on this journey, let’s continue

112
00:08:44,481 --> 00:08:46,920
our exploration of Machine Learning!

113
00:08:46,920 --> 00:08:48,450
Thanks for watching!