2. The Artificial Intelligence or AI is one of the Computer Science
area, which emphasizes the creation and development of
intelligent machines that can think, work and react like
humans.
CAPC3011 | Khaiziliyah Khalid
3. What is
Machine
Learning?
CAPC3011 | Khaiziliyah Khalid
Tom Mitchell (1998) - Machine Learning
is the study of algorithms that
improve their
performance
P
at some task
T
with
experience E
“Learning is any process by which a
system improves performance from
experience” – Herbert Simon
5. What is Machine Learning?
Follow instructions
Learn from experience
data
6. WHAT IS MACHINE LEARNING?
Machine Learning gives “computers the ability to learn without being explicitly programmed.”
(Samuel, A., 1959)
Machine Learning is the subset of Artificial Intelligence, that deal with the extraction of
patterns from data sets.
•This means that the machine can find rules for optimal behavior, but also can adapt to changes in the world
Because of new computing technologies, machine learning today is not like machine learning
of the past.
It was born from pattern recognition and the theory that computers can learn without being
programmed to perform specific tasks; researchers interested in artificial intelligence wanted
to see if computers could learn from data.
CAPC3011 | Khaiziliyah Khalid
8. History of Machine
Learning
CAPC3011
|
Khaiziliyah
Khalid
1970s:
• Symbolic concept induction
• Winston’s arch learner
• Expert systems and the knowledge acquisition bottleneck
• Quinlan’s ID3
• Michalski’s AQ and soybean diagnosis
• Scientific discovery with BACON
• Mathematical discovery with AM
9. History of Machine
Learning
CAPC3011
|
Khaiziliyah
Khalid
1980s:
• Advanced decision tree and rule learning
• Explanation-based Learning (EBL)
• Learning and planning and problem solving
• Utility problem
• Analogy
• Cognitive architectures
• Resurgence of neural networks (connectionism,
• backpropagation)
• Valiant’s PAC Learning Theory
• Focus on experimental methodology
10. History of Machine
Learning
CAPC3011
|
Khaiziliyah
Khalid
1990s
• Data mining
• Adaptive software agents and web applications
• Text learning
• Reinforcement learning (RL)
• Inductive Logic Programming (ILP)
• Ensembles: Bagging, Boosting, and Stacking
• Bayes Net learning
11. History of Machine
Learning
CAPC3011
|
Khaiziliyah
Khalid
2000s
•Support vector machines & kernel methods
•Graphica lmodels
•Statistical relational learning
•Transfer learning
•Sequence labeling
•Collective classification and structured outputs
•Computer Systems Applications (Compilers, Debugging,
•Graphics, Security)
•E-mail management
•Personalized assistants that learn
•Learning in robotics and vision
13. A classic example - It is very hard to
say what makes a 2
CAPC3011 | Khaiziliyah Khalid
14. Machine Learning Usage
CAPC3011 | Khaiziliyah Khalid
• Human expertise does not exist (navigating on Mars)
• Humans can’t explain their expertise (speech
recognition)
• Models must be customized (personalized medicine)
• Models are based on huge amounts of data (genomics)
ML is used when:
• There is no need to “learn” to calculate payroll
Learning isn’t always useful:
15. Machine Learning Usage
CAPC3011 | Khaiziliyah Khalid
• Facial identities or facial expressions
• Handwritten or spoken words
• Medical images
Recognizing patterns:
• Generating images or motion sequences
Generating patterns:
16. Machine Learning Usage
CAPC3011 | Khaiziliyah Khalid
• Unusual credit card transactions
• Unusual patterns of sensor readings in a
nuclear power plant
Recognizing anomalies:
• Future stock prices or currency exchange rates
Prediction:
18. Defining the Learning Task
T: Playing checkers
P: Percentage of games won against an arbitrary opponent
E: Playing practice games against itself
T: Recognizing hand-written words
P: Percentage of words correctly classified
E: Database of human-labeled images of handwritten words
T: Driving on four-lane highways using vision sensors
P: Average distance traveled before a human-judged error
E: A sequence of images and steering commands recorded while observing a
human driver.
T: Categorize email messages as spam or legitimate.
P: Percentage of email messages correctly classified.
E: Database of emails, some with human-given labels
19. Types of Learning
• Given: training data +
desired outputs (labels)
Supervised
learning
• Given: training data
(without desired outputs)
Unsupervised
learning
• Rewards from sequence
of actions
Reinforcement
learning
CAPC3011 | Khaiziliyah Khalid
20. Supervised
Learning
Definition: A method in which we teach the machine using labeled data
Problem type: Regression, Classification
Type of data: Labeled data
Training: External supervision
Aim: Forecast outcomes
Approach: Map labeled input to known output
Popular algorithms: Linear regression, logistic regression, support vector machine, k-nearest neighbor
Applications: Risk evaluation, forecast sales
CAPC3011 | Khaiziliyah Khalid
21. Unsupervised
Learning
Definition: The machine is trained on unlabeled data without any guidance
Problem type: Association, Clustering
Type of data: Unlabeled data
Training: No supervision
Aim: Discover underlying patterns
Approach: : Understand patterns and discover output
Popular algorithms: K-means, C-means, Apriori
Applications: Recommendation systems, anomalies detection
CAPC3011 | Khaiziliyah Khalid
22. Reinforcement
Learning
Definition: An agent interacts with its environment by producing actions or discovers errors or rewards
Type of data: No predefined data
Training: No supervision
Aim: Learn series of actions
Approach: Follow trial and error method
Popular algorithms: Q-learning, SARSA
Applications : Self-driving cars, gaming
CAPC3011 | Khaiziliyah Khalid
23. Example 1: Logistic Regression
CAPC3011 | Khaiziliyah Khalid
Price of a House
36. Machine
Learning in
Nutshell
CAPC3011 | Khaiziliyah Khalid
Tens of thousands
of machine
learning algorithms
Hundreds new every
year
Every ML algorithm
has three
components:
Representation: how
to represent
knowledge
Optimisation : the
way candidate
programs are
generated known as
the search process
Evaluation : the way
to evaluate candidate
programs
(hypotheses)
37. Various Function
Representations
CAPC3011 | Khaiziliyah Khalid
• Linear regression
• Neural networks
• Support vector
machines
Numerical
functions
• Decision trees
• Rules in
propositional logic
• Rules in first-order
predicate logic
Symbolic
functions
42. Characteristics of Machine Learning
Characteristics Explanation
The ability to perform
automated data
visualization
Machine learning offers several tools that provide rich snippets of
data which can be applied to both unstructured and structured data.
With the help of user-friendly automated data visualization platforms
in machine learning, businesses can obtain a wealth of new insights to
increase productivity in their processes.
Optimized to learn
complex patterns
Machine learning models are designed to be optimized to learn
complex patterns. In comparison to statistical models or decision tree
models, predictive models greatly excel, when you have very complex
patterns in data.
Account for interactions
and
nonlinear relationships
Machine learning predictive models can account for interactions in the
data and nonlinear relationships to an even better degree than
decision tree models.
Few assumptions These models are powerful because they have very few assumptions.
They can also be used with different types of data.
CAPC3011 | Khaiziliyah Khalid
43. The criteria needed while creating a good
machine learning systems
CAPC3011 | Khaiziliyah Khalid
Data preparation capabilities.
Algorithms – basic and advanced.
Automation and iterative processes.
Scalability.
Ensemble modelling.
44. Machine Learning in Practice
CAPC3011 | Khaiziliyah Khalid
Understand
domain, prior
knowledge, and
goals
Data integration,
selection, pre-
cleaning and
processing
Learning models
Interpreting
results
Consolidating and
deploying
discovered
knowledge
