Scribd
Upload
2 views13 pages

Machine Learning Concepts and Applications

The document outlines the fundamentals of Machine Learning, emphasizing its role in developing algorithms that allow computers to learn from data without explicit instructions. It discusses well-posed learning problems, which include defining tasks, experiences, and performance metrics, and provides examples of successful applications like speech recognition and self-driving cars. Additionally, it covers the design of learning systems, including choosing training experiences and target functions for effective learning.

Uploaded by

madhura gangaiah
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
2 views13 pages

Machine Learning Concepts and Applications

The document outlines the fundamentals of Machine Learning, emphasizing its role in developing algorithms that allow computers to learn from data without explicit instructions. It discusses well-posed learning problems, which include defining tasks, experiences, and performance metrics, and provides examples of successful applications like speech recognition and self-driving cars. Additionally, it covers the design of learning systems, including choosing training experiences and target functions for effective learning.

Uploaded by

madhura gangaiah
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd

Machine Learning II

MODULE 1

Dr. Madhura Gangaiah


Professor
Dept. of AI&ML
BGSCET
Subfield of AI that focuses on the development
Machine and study of the algorithms and statistical models
Learning that enable computers to perform tasks without
explicit instructions
Types of Machine Learning

This Photo by Unknown Author is licensed under CC BY-SA-NC


Successful Applications of
Machine Learning
• SPHINX system - Speech Recognition system
• ALVINN system – computer controlled vehicles to steer correctly
70 miles per hour for 90 miles
• TD – Gammon – playing world class gammon
Learned strategy by playing over one million practice games by itself
Well Posed Learning Problem
• Problem – task which needs to be solved
• Learning problem- problems that can be solved effectively using ML
• Well posed Learning Problem
Learning problem which has 3 features
• Task(T) – Objective

• Experience(E) – dataset from which algorithm can learn

• Performance(P) – metric used to assess the effectiveness of the algorithm at the task
Well Posed Learning Problem
A Computer program is said to learn from experience E with respect to some class of
tasks T and performance measure P, if its performance at tasks T, as measured by P,
improves with experience E

Checkers Game Robotic Driving Handwriting Recognition


Well Posed Learning Problem
Checkers Game
[Link]

ALVINN Self Driving Cars


[Link]
Well Posed Learning
Problem
Problem Task(T) Performance(P) Experience(E)
Playing Checkers Play against Make perfect It plays itself to
opponent to win moves to win the improve
game game
Handwriting Classify the Better Database of
Recognition images and text classification homework text
Robot Driving Drive car in a 4 Source to Images of
lane highway destination avg vehicles on road
distance travelled
Well Posed Learning Designing a
Problem Learning System

• Choosing the Training Experience


Type of training
• Direct
• Indirect
• Choosing the Target Function
• Choose representation of the target
function
Well Posed Learning Choosing the
Problem target function

1. Task T : playing Checkers


2. Performance Measure P: percent of games won in the world
tournament
3. Training Experience: games played against itself
4. Target function: V: Board -> R
5. Target function representation
V(b) = w +w x +w x +w x +w x +w x +w x
Well Posed Learning Choosing the
Problem target function

X1: the number of black pieces on the board


X2: the number of red pieces on the board
X3: the number of black kings on the board
X4: the number of red kings on the board
X5: the number of black pieces threatened by red
X6: the number of red pieces threatened by black
Choosing a function
Well Posed Learning approximation
Problem algorithm

• In order to learn the target function V, we require a set of training


examples, each describing a specific board state b and the training value
Vtrain(b)

• Each training example is ordered pair of the form (b,Vtrain(b)).

Example: Describes board state b in which black has won the game and for
which the target function value Vtrain(b) is +100.

((x1=3, x2 =0, x3=1, x4= 0, x5=0, x6=0),+100))


Well Posed Learning Estimating
Problem Training values

Only training information available to the learner is whether the game was
eventually won or lost.
We require training examples that assign specific scores to specific board states

Rule for estimating training values


Vtrain(b) <------ V(successor(b))

You might also like