UCCD2063

Artificial Intelligence Techniques

Unit 2:
The Fundamentals of Machine Learning
 Outline
• Machine Learning
Reference:
• Introduction
• Types of machine learning
• Challenges of Machine Learning
• The Machine Learning Framework

Géron Chapter 1
 Problems with Traditional Programming
Traditional programming paradigm:
Deploy

Study the
Write rules Evaluate
problem

Analyze errors

▪ Problems of traditional programming paradigm:

• complexity – too many rules, very hard to cover all aspects of the
problem, different problems require different rules
• static – cannot adapt to new input, need to keep writing new rules,
very hard to maintain

3
 The Machine Learning Framework
▪ Instead of handcraft rules, ML learns a model from training samples
(data)
▪ One learning algorithm for different problems

Training Deploy
Samples
Collect data

Study the Training ML Evaluate

problem algorithm Model Solution

Analyze errors

4
 What is Machine Learning?

“Machine Learning: Field of study

that gives computers the ability to
learn (from data) without being
explicitly programmed. ”
[Arthur Samuel (1959)]

5
 Why use Machine Learning?
▪ Some problems are too complex to solve by using rules
• An example: image classification

Input Complex, specific Output

program “car”
if this local region looks like a side mirror,
and
if this local region area looks like a wheel
then classify image as a car

Will work if given the same image

again, but, given new images, the
algorithm is bound to fail

6
 Why use machine learning?

▪ Good for problems that evolves with time

• Machine learning can automatically rebuild the model when
necessary
• Example: spam classifier – learns new spam words when it become
unusually frequent in spam flagged by users

Training Update
Deploy
Samples data

Can be
updated
Study the Train ML Evaluate
problem algorithm Solution

Analyze errors

7
 Why use machine learning?

▪ Help humans learn

• Can inspect ML to see what they learn
− may find unsuspected correlations or new trends
− learn the problem better
− Example: can examine the list of words or its combination that ML identifies as the
best predictors of spam filter

Lots of
training samples

Study the Train ML

Solution
problem algorithm

Iterate if Understand the Inspect the

necessary problem better solution

8
 Data in Machine Learning
▪ Numerical (quantitative) data
• Discrete (e.g. 1, 3, 8, -4, …)
• Continue (e.g. 2.321, 0.2437, …)
▪ Categorical (qualitative) data
• Ordinal (e.g. low, medium, high)
• Nominal (e.g. red, blue, yellow)

9
 Data in Machine Learning
▪ Labelled data vs unlabelled data
• Labelled data: data that comes with a tag (e.g. name, value)
• Unlabelled data: data that comes with no tag

10
 Structured Data vs Unstructured Data

Structured Data
• Specific and stored in a
predefined format
• Suitable for traditional
machine learning
• Focused in this course

Unstructured
Data
• Collection of varied types of
data that are stored in their
native formats (e.g. text,
image, video, audio,…)
• Better result with deep
learning techniques 11
Type of Machine Learning Algorithms

Labeled Unlabeled Mixed No Training

Data Data Data Data

12
 Supervised Learning

▪ In supervised learning, the algorithm is given some example

input-output pair and it learns a function that maps from input
to output
▪ input: the set of features used to describe the samples
▪ output: the attribute we are interested to predict
▪ Example: Fruit classification
Input Output
Features
length width weight Label
fruit 1 165 38 172 Banana
fruit 2 218 39 230 Function Banana
Samples
fruit 3 76 80 145 Orange
fruit 4 145 35 150 Banana
… … … … …
fruit m … … … …

13
 Supervised Learning Tasks

Two main supervised learning tasks

Classification Regression
Classification predicts discrete Regression predicts continuous
valued output (e.g., present/not valued output (e.g., house price)
present) 400

300
Price (RM x1000)
Yes No 200

100

0 Size
0 500 1000 1500 2000 2500
Object Detection (Images with Car)
Housing Price Prediction

14
Supervised Learning – Classification Applications
Income classification:
• Input: numerical and categorical data
• Output: 1 (income<=50K), 0 (income>50K) – discrete
• Features: age, workclass, marital-status, race, education, area, …

15
 More Classification Applications

Digit Classification Input (Image): Output (discrete):

▪ Input: images / pixel grids 0
▪ Output: a digit 0-9
Feature 1
▪ Features:
Extraction
• Signatures
2
• Histogram of gradients
• Shape Patterns: NumComponents, 1
AspectRatio, NumLoops
• … ??

Spam mail classification Input (Text): Output (discrete):

▪ Input: an email Dear Andy,
▪ Output: spam or non-spam How you are doing, buddy? Hopefully you
are adapting well to your new school. All
▪ Features: of us miss you dearly here. We miss your
silly jokes.
• Words: FREE!, Earn, Call now,..
• Text Patterns: $$$, ALL CAPS Hello, I have a special offer for you…
WANT TO LOSE WEIGHT? The most
• Non-text: SenderInContacts powerful weight loss is now available
• … without prescription.

16
Supervised Learning – Regression Applications
Predict the house price in the Boston area (regression):
• Input: numerical and categorical data
• Output: house price (in 1000usd) – continue
• Features:
− CRIM: per capita crime rate by town
− ZN: proportion of residential land
− INDUS: proportion of non-retail business acres per town
− CHAS: Charles River dummy variable (= 1 if near river; 0 otherwise)
− NOX: nitric oxides concentration (parts per 10 million)
− RM: average number of rooms per dwelling
− AGE: proportion of units built prior to 1940
− DIS: weighted distance to employment centres
− RAD: index of accessibility to radial highways
− TAX: full-value property-tax rate per $10,000
− …

17
 A Simple Supervised Learning Example (1/6)

▪ Problem: want to predict IT salary

▪ Step 1: Collect Data
• Consult domain expert/survey/research what key factors
(features) affecting IT salary
− Experience in years (used in this example)
− Job title
− Size of organization
Samples
− Gender X (year) y (salary, k)
− Industry sector 5 38
10 64
− Geographic region 7 36
− ... 8 44
0 na
• Collect the data … …
− From survey/database
− Data processing for missing/incomplete data, normalization (see L03)

18
 A Simple Supervised Learning Example (2/6)

▪ Step 2: Model selection

• Study the data and determine what model is suitable
− For this plot, select linear regression
𝑦 = 𝜃0 + 𝜃1 × 𝑥 𝜃0 , 𝜃1 = model parameters

y
(salary)

X (year)

ChatGPT has 175 billion parameters, GPT4 > 1 trillion parameters 19

 A Simple Supervised Learning Example (3/6)

▪ Step 3: Train model

• Separate data into training set (80%) and test set (20%)
− Train model on training set, validate model using test set

Training set
Test set

y
(salary)

X (year)

20
 A Simple Supervised Learning Example (4/6)

▪ Step 3: Train model (cont.)

• Train model with training set
− Use normal equation or gradient descent (see L05)
− Minimize sum-of-squared error (SSE) – training error

𝑦 = 𝜃෠0 + 𝜃෠1 𝑥
prediction error

y
(salary)

X (year)
21
 A Simple Supervised Learning Example (5/6)

▪ Step 4: Validate model

• Validate the model using test set – test error

y
(salary)

X (year)

22
 A Simple Supervised Learning Example (6/6)

▪ Step 5: Deploy model

• Use the trained model for prediction

𝑦 = 𝜃෠0 + 𝜃෠1 𝑥
y
(salary)

X (year)

23
 Supervised Learning Algorithms

▪ Algorithms for classification:

• k-Nearest neighbour (k-NN)
• Logistic Regression
• Decision Tree and Random Forests
• Support Vector Machine (SVM)
• Neural Networks (NN)
• ...

▪ Algorithms for regression:

• K-NN Regressor
• Linear Regression
• Decision Tree and Random Forests
• SVM Regressor
• NN
• Non-linear Regression ...
24
 Unsupervised learning

Unsupervised Learning Features

▪ No labels are provided for all training length width weight
fruit 1 165 38 172
samples fruit 2 218 39 230
▪ Discovers the underlying structure, fruit 3 76 80 145
fruit 4 145 35 150
relationship or patterns based only on … … … …
the features of the training sample fruit m … … …

Supervised Learning Unsupervised Learning

Positive
samples cluster
x x
x
vs
x2 x
x x2
Negative Decision Samples (no labels)
samples Boundary

x1
x1 25
 Unsupervised Learning Task: Clustering

▪ Detect groups of similar samples

▪ Example:
Detecting groups of visitors who visit your blog
Example analysis:
Time of day • 40% visitors who love comic
books and read in the evening
night • 20% are young sci-fi lovers who
visit before school, etc.

day How does it help?

Age • Can target your posts for each
young -> old group

26
Unsupervised Learning Task: Anomaly detection
▪ Identify items, events or observations which do not conform
to an expected pattern or other items in a dataset
▪ Anomalies are also referred to as outliers, novelties or
noise

outliers

Applications: Bank fraud, medical problems or errors in a text. 27

 Unsupervised Learning Task: Association Rule Mining

▪ Given a set of transactions, find rules that will predict the

occurrence of an item based on the occurrences of other items
in the transaction.

Support(Bread) = #Bread / #total

Support(Jam) = #Jam / #total
Support(Bread, Jam) = #(Bread+Jam) / #total

Confident(Bread->Jam) = #(Bread+Jam) / #Bread

Lift(Bread->Jam) = Support(Bread, Jam) /
(Support(Bread) x Support(Jam))

28
 Unsupervised Learning Tasks and Algorithms

▪ Clustering
• k-Means
• Hierarchical Cluster Analysis (HCA)
• Expectation Maximization

▪ Association rule learning

• Apriori
• Eclat

29
 Semi-supervised learning
Features
▪ Partially labeled training data. length width weight
Label

Typically more unlabeled data fruit 1 165 38 172 Banana

fruit 2 218 39 230 ?
than labeled fruit 3 76 80 145 Orange
fruit 4 145 35 150 ?
… … … … …
▪ Most semi-supervised fruit m … … … …

algorithms are combinations

of unsupervised algorithms
and supervised algorithms

30
 Reinforcement Learning

▪ No training set is provided

▪ Learns based on the feedback of the environment:
• effect of the agent' action on the environment (state)
• rewards of taking a particular action
▪ Learns by itself the best policy (state-action) to get the
most reward over time
Policy

Applications: games (chess, go, video), robotics, traffic control, trading,... 31

Reinforcement Learning Example

The agent
1. observes the environment
2. Select action using policy

3. Perform action
4. Get reward or penalty

5. Update policy (learning step)

6. Iterate until an optimal policy
is found

32
 Differences between Machine Learning Types

Supervised Learning Unsupervised Learning Reinforcement Learning

Labeled data with output Unlabeled data, output not Environment with rewards and
specified specified penalty

Solves problems by mapping Solves problems by discovering Solves problems by trial and
input to known output underlying patterns error

External supervision No supervision No supervision

Used for regression and Used for clustering and Used for control and decision
classification tasks association tasks making tasks

33
 Foundation Model & Self-Supervised Learning
▪ Conventionally, a AI model is trained on task-specific data to
perform very specific task.
▪ A new paradigm in AI has emerged called foundation models.
Unlike traditional AI, foundation models learn from massive
datasets across different domains.
▪ Through self-supervised learning techniques, a foundation
model teach itself to acquire broad scope of knowledge and
understanding of the world (general intelligence).
▪ Large language models (LLM) such as OpenAI's GPT-4 and
Google's PaLM are examples of foundation models .
▪ The foundation models can then be transferred to perform any
other tasks through fine-tuning or prompting.
• GPT -> ChatGPT, GPT -> Copilot, GPT -> Duolingo
▪ Foundation models require a lot more data and computing
power to train.
34
 Self-Supervised Learning
▪ Self-supervised learning is a new machine learning process
where the model trains itself to learn one part of the input
from another part of the input to obtain useful representations
and knowledge.
▪ The trained model can help with downstream learning tasks.

Random shaffle Ground truth

35
 How LLMs are trained

Self-supervised learning
Source: Borealis AI 36
 Challenges of Machine Learning

▪ Insufficient quantity of training data

▪ Poor data quality
▪ Non-representative training data
▪ Irrelevant features
▪ Underfitting & overfitting the training data

37
 Poor-quality data
▪ Training data may contain errors, for example:

Outlier

▪ Data cleaning: Most data scientists spend a significant time to clean the
data. For example:
• Fill up missing value
• Drop a column (feature) with many missing values/errors
• Remove rows (samples) with outliers
• Fix error/format manually

38
 Non-representative data

▪ Training data should be representative of the new cases that you want
to generalize to.
▪ Consider fitting a linear model to the GDP dataset with and without 7
missing countries :
(This model does not generalize well)
without missing
countries

with missing
countries

39
 Irrelevant features
▪ Selected features must be relevant to the task at hand. Having
irrelevant features in your data can decrease the accuracy of
the models.
• For example, area or perimeter length are irrelevant feature for
classifying shapes like circle and rectangle.

f(area) = ?
• Features such as signature, number of corners are more suitable for
classifying shapes.

40
 Feature Engineering

▪ Feature engineering is the process to come up with a

good set of features to train on. It involves:
• feature extraction – use some tools to extract features
from samples (e.g., extract shape signature, colors
from an object).
• feature selection – choose the most useful features
among all existing features that produce the best
result for a machine learning model.

Deep learning learns features automatically but requires lots of training data.

41
 Underfitting & Overfitting
▪ Underfitting (high bias) may happen when our model is over-simplified
or not expressive enough (high training error and high test error).

▪ Overfitting (high variance) may happen when our model is too complex
and fits too specifically to the training set, but it does not generalize
well to new data (low training error but high test error).

= test data 42
 Underfitting & Overfitting
▪ Underfitting and overfitting in regression task

43
 Hyperparameter Tuning

▪ Hyperparameter is a parameter whose value is set before the learning

process begins and which is used to control the learning process.
• For example, % of train-test split is a hyperparameter
▪ Different model has different set of hyperparameters. For example,
• Polynomial model: degree
• K-NN: n_neighbors, distance metric (Manhattan or Euclidean)
• Neural Networks: α (learning rate), max_iter (maximum iterations)
• SVM: kernel (linear, rbf), C (penalty parameter)
▪ In machine learning, hyperparameter tuning is the process of choosing
a set of optimal hyperparameters for a learning algorithm.
▪ Need to balance between fitting the data perfectly and keeping the
model simple to ensure it generalizes well (avoid underfit & overfit)
▪ Hyperparameter tuning is an important step of building a machine
learning system.
44
 The Machine Learning Framework
▪ Divided into two phases
Training Phase: Testing Phase:

Raw Data Data Preprocessing

(cleaning, normalization, dimension reduction, etc.)

Training Validation Testing

Set Set Set

Learning Test
Hyperparameter
Algorithm
Tuning & Tune
Validation

Train Model Final Model

Predict

45
Next:

The Regression Pipeline