Introduction to

Machine Learning
LÊ ANH CƯỜNG
Ton Duc Thang University
2024

1
Content
1. What is AI and what is ML?
2. How ML helps or why learn ML?
3. The general diagram of machine learning model building; what is
model in ML?
4. Key components of ML
5. Type of machine learning methods
6. History of ML

2
Content
1. What is AI and what is ML?
2. How ML helps or why learn ML?
3. The general diagram of machine learning model building
4. Key components of ML
5. Type of machine learning methods
6. History of ML

3
AI and ML
• What is Artificial Intelligence (AI)?
• What is Machine Learning (ML)?
• Relationship between AI and ML?

4
AI
• Artificial Intelligence (AI) is a field of study focused on methods and
techniques that enable computers and machines to simulate human
intelligence, demonstrated through capabilities such as:
• human learning,
• comprehension,
• problem solving,
• decision making,
• creativity and autonomy.

• Trí tuệ nhân tạo (AI) là lĩnh vực nghiên cứu phương pháp, kỹ thuật cho
phép máy tính và máy móc mô phỏng khả năng học tập, hiểu biết, giải
quyết vấn đề, ra quyết định, sáng tạo và tự chủ của con người.

5
Machine Learning
• Machine learning involves creating models by training an algorithm to
make predictions or decisions based on data.
• It encompasses a broad range of techniques that enable computers to learn
from and make inferences based on data without being explicitly
programmed for specific tasks.

• Học máy liên quan đến việc tạo ra các mô hình bằng cách đào tạo một thuật
toán để đưa ra dự đoán hoặc quyết định dựa trên dữ liệu.
• Nó bao gồm một loạt các kỹ thuật cho phép máy tính học hỏi và đưa ra suy
luận dựa trên dữ liệu mà không cần được lập trình rõ ràng cho các tác vụ cụ
thể.

6
Narrow AI

• AI is expected as experts:
• AI offers optimal solutions to difficult/complicated problems
• AI solves predictive and focasting problems (uncertainty)
• Interacts like human

correct

question AI answer
Interacts Interacts
like human like human

7
 What is Machine Learning?

Data
Machine
(Examples) Learning

The optimal
Input of the solution
problem

question AI model answer

Interacts Interacts
like human like human

8
AI vs ML?

9
https://www.ibm.com/topics/artificial-intelligence
10
ML vs explicit Algorithms?
• ML: It encompasses a broad range of techniques that enable computers
to learn from and make inferences based on data without being
explicitly programmed for specific tasks
• ML is a black box model, it can not explain clearly (why, meaning) the
result is generated.

11
Content
1. What is AI and what is ML?
2. How ML helps or why learn ML?
3. The general diagram of machine learning model building
4. Key components of ML
5. Type of machine learning methods
6. History of ML

12
13
Domains of applying AI
• Retail
• AI provides virtual shopping capabilities that offer personalized
recommendations and discuss purchase options with the consumer.
• Banking
• Artificial Intelligence enhances the speed, precision and effectiveness of
human efforts. In financial institutions, AI techniques can be used to identify
which transactions are likely to be fraudulent, adopt fast and accurate credit
scoring, as well as automate manually intense data management tasks.
• Games
• AI in games is about planning. The artificial agent must decide on the next
best action(s) to perform depending on the state of the game.
14
Domains of applying AI
• Transportation
• autonomous cars; navigation system; monitoring traffic lights;
• Logistics and Warehouse Management
• Logistics and warehouses management are two very complex problems.
They depends on many parameters difficult to model, some of them even time-
dependent.
• Machine Translation
• Virtual Assistant
• Computer Vision
• Healthcare
• …
15
Content
1. What is AI and what is ML?
2. How ML helps or why learn ML?
3. The general diagram of machine learning model building
4. Key components of ML
5. Type of machine learning methods
6. History of ML

16
How does an Expert solve problems?

Problem Solution
(question) Expert (answer)
Expert
Input

Knowledge Inference

Output

18
Expert
Input

Experience Knowledge
Acquisition Knowledge Reasoning/Inference
(Data),
Teachers (Learning)

Output

19
A General Paradigm of ML Systems
input

ML algorithms Knowledge Reasoning/Inference

Data

model output

20
Content
1. What is AI and what is ML?
2. How ML helps or why learn ML?
3. The general diagram of machine learning model building
4. Key components of ML
5. Type of machine learning methods
6. History of ML
7. Machine learning and other fields

21
What are ML components in AI systems?
• Data
• Model
• Learning algorithms

22
An example of a machine learning problem

23
An example of a machine learning problem

24
An example of a machine learning problem

Model Learning
Model
selection (training, fitting)

data 25
An example of a machine learning problem

y = a*x + b

Weight = a*Height + b

Model Learning
Model
selection (training, fitting)

data 26
Weight = w1*Height + w0

27
How to understand Learning?
=> learn values for model’s parameters by inferencing
from given examples Prediction
model

y = ax + b Learning y=2*x + 5

Model
Selection
Examples
{(xi,yi)}n
28
ML model:
example 2 Decision Tree

29
ML model:
example 3

30
A General Machine Learning Diagram
Data
New input
Representation

Data Mode Learning Learnt

Data Representation Method Model

output
Model
Selection

SVM, NB, DT, LR, NN, kNN,…

31
What should be learnt in Machine Learning?
• Understanding ML models
• Architecture, Structure, or form of the models
• Computation, inference/reasoning
• Methods/Algorithms for learning models’ parameters
• Model evaluation
• Data Processing: data collection; feature selection; dimensional
reduction, noisy filtering,…
• Other issues: overfitting,

32
Content
1. What is AI and what is ML?
2. How ML helps or why learn ML?
3. The general diagram of machine learning model building
4. Key components of ML
5. Type of machine learning methods
6. History of ML

33
Types of Machine Learning

34
Content
1. What is AI and what is ML?
2. How ML helps or why learn ML?
3. The general diagram of machine learning model building
4. Key components of ML
5. Type of machine learning methods
6. History of ML

35
History of ML

36
History of ML
Generative Model
(Very complicate)
Recognition ChatGPT,
(complicated) Gemini,…
Recognition (Transformers,
CNN, LSTM, GAN,…
(Simple) (Bigdata, GPU)
Very Large
models)
Perceptron, MLP,
NB, LR, SVM

1950 1980 2000 2010 2022

thousands of parameters millions of parameters billions of parameters

37
Topics of this course
1. Machine Learning: an 5. Ensemble Learning
Overview 6. Unsupervised Learning
2. Classification 1. Clustering
1. NB classifiers, Generative vs 2. Dimension Reduction
Discriminative models
2. Decision Tree and Random
Forest
7. Neural Networks
3. Support Vector Machine 1. Feed Forward Neural
4. K-Nearest Neighbor Networks
2. Recurrent Neural Network
3. Regression
1. Linear Regression and 8. Advanced Topics
Gradient Descent Algorithm 1. Deep Learning
2. Logistic Regression for
classification 2. Reinforcement Learning
4. Overfitting

38
Course assessment
• Progress exercises (30%)
• Middle exam (20%)
• Project and presentation
• Final exam (50%)
• Project and presentation

39