Download Machine Learning

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts
no text concepts found
Transcript
Machine Learning
1. Where does machine learning fit in computer science?
2. What is machine learning?
3. Where can machine learning be applied?
4. Should I care about machine learning at all?
1
Where does machine learning fit in computer
science?
Field of Study
Artificial Intelligence
Planning
Knowledge
Representation
Search
Machine Learning
Robotics
Clustering
Reinforcement
Learning
Genetic Algorithms
Classification
2
Where does machine learning fit in computer
science? (2)
Multidisciplinary Field
Probability &
Statistics
Computational
Complexity
Theory
Artificial
Intelligence
Neurobiology
Machine
Learning
Information
Theory
Philosophy
3
Where does machine learning fit in
computer science? (3)
Knowledge Discovery and Data Mining
Selection
Data
Preprocessing
Target Data
Knowledge
Transformation
Preprocessed
Data
Interpretation &
Evaluation
Patterns
Transformed
Data
Data
Mining
4
Machine Learning
1. Where does machine learning fit in computer science?
2. What is machine learning?
3. Where can machine learning be applied?
4. Should I care about machine learning at all?
5
Machine Learning
•
Where does machine learning fit in computer science?
•
What is machine learning?
•
Definition
•
Types of Machine Learning
•
Where can machine learning be applied?
•
Should I care about machine learning at all?
6
What is Machine Learning?
Definition
Machine learning is the study of how to make computers
learn or adapt; the goal is to make computers improve their
performance through experience.
Class of Tasks T
Computer
Performance P
Learning
Algorithm
Experience E
7
What is Machine Learning?
Definition (2)
Class of Tasks T
Computer
Performance P
Learning
Algorithm
Experience E
8
What is Machine Learning?
Definition (3)
Class of Tasks:
It is the kind of activity on which the computer will learn to
improve its performance. Examples:
Learning to
Play chess
Recognizing
Images of
Handwritten
Words
Diagnosing
patients
coming into the
hospital
9
What is Machine Learning?
Definition (4)
Class of Tasks T
Computer
Performance P
Learning
Algorithm
Experience E
10
What is Machine Learning?
Definition (5)
Experience and Performance
Experience: What has been recorded in the past
Performance: A measure of the quality of the response or action.
Example:
Handwritten recognition using Neural Networks
Experience: a database of handwritten images
with their correct classification
Performance: Accuracy in classifications
11
What is Machine Learning?
Definition (6)
Class of Tasks T
Computer
Performance P
Learning
Algorithm
Experience E
12
What is Machine Learning?
Definition (7)
Designing a Learning System
The Representation of the Target Knowledge
Example: Diagnosing a patient coming into the hospital.
Features:
 X1: Temperature
 X2: Blood pressure
 X3: Blood type
 X4: Age
 X5: Weight
 Etc.
Given a new example X = < x1, x2, …, xn >
F(X) = w1x1 + w2x2 + w3x3 = … + wnxn
If F(X) > T predict heart disease
otherwise predict no heart disease
13
Machine Learning
•
Where does machine learning fit in computer science?
•
What is machine learning?
•
•
Definition
•
Types of Machine Learning
•
Where can machine learning be applied?
Should I care about machine learning at all?
14
What is Machine Learning?
Types of Machine Learning
• Supervised Learning
• Unsupervised Learning
• Reinforcement Learning
• Evolutionary Learning
15
What is Machine Learning?
Types of Machine Learning (2)
Supervised Learning
Each example or object has a class attached to it.
We try to learn a mapping from examples to classes.
Two modes: classification and regression
Machine learning algorithms abound:
 Decision Trees
 Rule-based systems
 Neural networks
 Nearest-neighbor
 Support-Vector Machines
 Bayesian Methods
16
What is Machine Learning?
Types of Machine Learning (2)
Supervised Learning – Neural Networks
Neuron
Model Representation
Brain
Artificial Neural Networks are crude attempts to model the highly massive
parallel and distributed processing we believe takes place in the brain.
Consider:
1) the speed at which the brain recognizes images;
2) the many neurons populating a brain;
3) the speed at which a single neuron transmits signals.
17
What is Machine Learning?
Types of Machine Learning (2)
Supervised Learning – Neural Networks(2)
Left
Straight
Right
Output nodes
Internal nodes
Input nodes
18
What is Machine Learning?
Types of Machine Learning (3)
Unsupervised Learning
Examples or objects have no class attached to them.
From “Pattern Classification” by Duda, Hart and Stork, 2nd Ed. Wiley Interscience (2000)
19
What is Machine Learning?
Types of Machine Learning (4)
Reinforcement Learning
Supervised Learning:
Example
Class
Reinforcement Learning:
…
Situation Reward
Situation Reward
20
What is Machine Learning?
Types of Machine Learning (5)
Evolutionary Learning
Methods inspired by the process of biological evolution.
Main ideas
Population of
solutions
Generate new
solutions (offspring)
Assign a score or
fitness value to each
solution
Retain the best
solutions (survival
of the fittest)
21
Machine Learning
•
Where does machine learning fit in computer science?
•
What is machine learning?
•
Where can machine learning be applied?
•
Should I care about machine learning at all?
22
Where can machine learning be applied?
Application 1
Automatic car drive (ALVINN 1989)
Train computer-controlled vehicle to steer correctly when
driving on a variety of road types.
computer
(learning algorithm)
class 2
class 1
steer to the right
steer to the left
class 3
continue straight
23
Where can machine learning be applied?
Application 1 (2)
Automatic Car Drive
Class of Tasks:
Learning to drive on highways from
vision stereos.
Knowledge:
Images and steering commands recorded
while observing a human driver.
Performance Module: Accuracy in classification
24
Where can machine learning be applied?
Application 2
Learning to classify astronomical structures.
galaxy
stars
Features:
o Color
o Size
o Mass
o Temperature
o Luminosity
unkown
25
Where can machine learning be applied?
Application 2 (2)
Classifying Astronomical Objects
Class of Tasks:
Learning to classify new objects.
Knowledge:
database of images with correct
classification.
Performance Module: Accuracy in classification
26
Where can machine learning be applied?
Other Applications
 Bio-Technology
 Protein Folding Prediction
 Micro-array gene expression
 Computer Systems Performance Prediction
 Banking Applications
Credit Applications
 Fraud Detection
 Character Recognition (US Postal Service)
 Web Applications
 Document Classification
 Learning User Preferences
27
Machine Learning
•
Where does machine learning fit in computer science?
•
What is machine learning?
•
Where can machine learning be applied?
•
Should I care about machine learning at all?
28
Should I care about Machine Learning at all?

Yes, you should!

Machine learning is becoming increasingly popular and has become a
cornerstone in many industrial applications.

Machine learning provides algorithms for data mining, where the goal is
to extract useful pieces of information (i.e., patterns) from large
databases.

The computer industry is heading towards systems that will be able to
adapt and heal themselves automatically.

The electronic game industry is now focusing on games where
characters adapt and learn through time.

NASA is interested in robots able to adapt in any environment
autonomously.
29
Summary

Machine learning is the study of how to make computers learn.

A learning algorithm needs the following elements: class of tasks,
performance metric, and body of experience.

The design of a learning algorithm requires to define the knowledge to
learn, the representation of the target knowledge, and the learning
mechanism.

Machine learning counts with many successful applications and is
becoming increasingly important in science and industry.
30