DAT21003
Artificial Intelligence
Basic Problem Solving in Machine Learning
Introduction
Learning in Artificial Intelligence
Learning is the ability of an intelligent agent, to improve
its behaviour, based on experience.
This means:
The range of behaviours / tasks is expanded (the agent can
do more).
The accuracy of tasks is improved (the agent can do things
better).
The speed is improved (the agent can do things faster).
Introduction
Source: What is Machine Learning?
Introduction
Machine Learning
Machine learning is the field of study, that gives
computers (machines) the capability to learn, without
being explicitly programmed.
Machine learning algorithms build a mathematical model
based on sample data (known as training data).
This mathematical model, enables the machine to
perform tasks (make predictions / decisions), without
explicitly programmed to do so.
Learning Tasks
Why Machine Needs to ‘Learn’?
Human expertise does not exist.
Example: navigating on Mars.
Humans are unable to explain their expertise.
Example: speech recognition.
Solution changes in time.
Example: routing on a computer network.
Solution needs to be adapted to cases.
Example: user biometrics.
Types of Machine Learning Algorithm
Machine learning algorithms can be traditionally
categorized into three broad categories:
Supervised Unsupervised Reinforcement
Learning Learning Learning
Types of Machine Learning Algorithm
1. Supervised Learning
Supervised learning algorithms build a mathematical
model of a set of training data.
The training data contains both the inputs and the desired
outputs.
The goal is to learn a general rule that maps inputs to
outputs.
The training process continues until the model achieves
the desired level of accuracy on the training data.
Types of Machine Learning Algorithm
1. Supervised Learning
Types of Machine Learning Algorithm
1. Supervised Learning
Most common algorithms:
Support Vector Logistic Naive Bayes
Linear Regression
Machine (SVM) Regression Classifier
Artificial Neural
Linear k-Nearest
Network (ANN) –
Discriminant Decision Trees Neighbours
Multilayer
Analysis (LDA) (k-NN)
Perceptron (MLP)
Types of Machine Learning Algorithm
1. Supervised Learning
Example application for classification:
Credit scoring
Types of Machine Learning Algorithm
1. Supervised Learning
Example application for classification:
Face recognition
Types of Machine Learning Algorithm
1. Supervised Learning
Example application for classification:
Character recognition – identify handwriting style
Types of Machine Learning Algorithm
1. Supervised Learning
Example application for classification:
Speech recognition
Sound Wave
Types of Machine Learning Algorithm
1. Supervised Learning
Example application for classification:
Gesture recognition
Types of Machine Learning Algorithm
1. Supervised Learning
Example application for classification:
Medical diagnosis
Types of Machine Learning Algorithm
1. Supervised Learning
Example application for classification:
Brainwaves interpretation
Types of Machine Learning Algorithm
1. Supervised Learning
Example application for classification:
Reading text
Types of Machine Learning Algorithm
1. Supervised Learning
Example application for regression:
Used-car price prediction
Types of Machine Learning Algorithm
1. Supervised Learning
Example application for regression:
Kinematics of a robotic arm
Types of Machine Learning Algorithm
2. Unsupervised Learning
Unsupervised learning algorithms take a set of data that
contains only inputs, and find structure in the data, like
grouping / clustering of data points.
The algorithms, therefore, learn from test data that has
not been labelled, classified / categorized.
Types of Machine Learning Algorithm
2. Unsupervised Learning
Types of Machine Learning Algorithm
2. Unsupervised Learning
Most common algorithms:
Principal Singular Value Independent
k-Means
Component Decomposition Component
Clustering
Analysis (SVD) Analysis
Types of Machine Learning Algorithm
2. Unsupervised Learning
Example application for clustering:
Customer segmentation – company applies different
marketing strategy for different customer groups / clusters.
Customers who watched
this item (The Imitation
Game) also watched …
Types of Machine Learning Algorithm
2. Unsupervised Learning
Example application for anomaly detection:
Fraud detection – system is trained with a lot of normal
instances; so, when it sees an unusual instance, it can
detect whether it is an anomaly / not.
Types of Machine Learning Algorithm
3. Reinforcement Learning
Reinforcement learning is a learning method that interacts
with its environment by producing actions and discovers
errors / rewards.
Trial and error search, and delayed reward are the most
relevant characteristics of reinforcement learning.
Types of Machine Learning Algorithm
3. Reinforcement Learning
Types of Machine Learning Algorithm
3. Reinforcement Learning
This method allows machines and software agents to
automatically determine the ideal behaviour within a
specific context to maximize its performance.
Simple reward feedback is required for the agent to learn
which action is best; this is known as the reinforcement
signal.
Types of Machine Learning Algorithm
3. Reinforcement Learning
Example application:
Play games
Deep Blue AlphaGo
