Está en la página 1de 24

m  


 

‡ 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?
Ê 


  

  
  

   
à 



  

  
    

 
 

  

 
 Ê 
Artificial
Intelligence

Probability & Neurobiology


Statistics
Õachine
Learning

Computational
Complexity Philosophy
Theory Information
Theory
j   
    

  j         




j  j     

   ! 


m   
 

‡ Where does machine learning fit in computer science?


‡ What is machine learning?
‡ Definition
‡ Design of a learning system
‡ Where can machine learning be applied?
‡ Should I care about machine learning at all?
 

Õachine learning is the study of how to make computers


learn; the goal is to make computers improve their
performance through experience.

Class of Tasks Computer Performance 


Learning
Algorithm

Experience =
`j"

Class of Tasks Computer Performance 


Learning
Algorithm

Experience =
`j"

It is the kind of activity on which the computer will learn to


improve its performance. Examples:

 
r   
 



   
 
 

!#     

Class of Tasks Computer Performance 


Learning
Algorithm

Experience =
!#     

=  
: What has been recorded in the past

 
: A measure of the quality of the response or action.

= 

Handwritten recognition using Neural Networks

= :
: a database of handwritten images
with their correct classification
 
:: Accuracy in classifications
 

m   
 

‡ Where does machine learning fit in computer science?


‡ What is machine learning?
‡ Definition
‡ Design of a learning system
‡ Where can machine learning be applied?
‡ Should I care about machine learning at all?
   
   

Class of Tasks Computer Performance 


Learning
Algorithm

Experience =
j $    j  

Example: Diagnosing a patient coming into the hospital.

Features:
 0 : Temperature
 0: Blood pressure
 0: Blood type
 0: Age Given a new example 0 = < x  x « xn >
 0: Weight
 Etc. F(0 = w x + wx + wx = « + wnxn

If F(0 > T predict heart disease


otherwise predict no heart disease
j $    j  

There are many possibilities:


[ The class of functions is very expressive.
You can represent almost any function but to be effective
the method needs lots of examples.
[ The class of functions is very limited.
Don¶t need many examples but may fail to contain the
true target function.
j 
    

Õachine learning algorithms abound:


[ Decision Trees
[ Rule
Rule--based systems
[ Neural networks
[ Nearest
Nearest--neighbor
[ Support
Support--Vector Õachines
[ Bayesian Õethods

Important characteristics of the learning mechanism:


‡ What is the class of functions
‡ How do you search over the class of functions
m   
 

‡ 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?
Ô%





!"#"$

%
 &   
 

 
  '





 '


' (



 
  
 $


 
!  
)
 


  


 


 
*
  
 
Ô%

Ô  

`    :
   : Learning to drive on highways from
vision stereos.
  :
: Images and steering commands recorded
while observing a human driver.
 
   :: Accuracy in classification
Ô&

Learning to classify astronomical structures.

galaxy

stars

Features:
o Color
o Size
o Õass
o Temperature unkown
o Luminosity
Ô&

 Ô   


`    :
   : Learning to classify new objects.
  :
: database of images with correct
classification.
 
   :: Accuracy in classification
Î Ô

ü Bio
Bio--Technology
ü Protein Folding Prediction
ü Õicro
Õicro--array gene expression
ü Computer Systems Performance Prediction
ü Banking Applications
üCredit Applications
ü Fraud Detection
ü Character Recognition (US Postal Service
ü Web Applications
ü Document Classification
ü Learning User Preferences
m   
 

‡ 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?
  '  
  

 Yes you should!


 Õachine learning is becoming increasingly popular and has become a
cornerstone in many industrial applications.
 Õachine 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.
 

[ Õachine learning is the study of how to make computers learn.


[ A learning algorithm needs the following elements:
   
 
 
    

[ The design of a learning algorithm requires to define the   
          

 
[ Õachine learning counts with many successful applications and is
becoming increasingly important in science and industry.

También podría gustarte