MACHINE
LEARNING
UNIT - 1
INTRODUCTION TO MACHINE LEARNING
Review of Linear Algebra for machine learning;
Introduction and motivation for machine learning;
Examples of machine learning applications, Vapnik-
Chervonenkis (VC) dimension, Probably Approximately
Correct (PAC) learning, Hypothesis spaces, Inductive
bias, Generalization, Bias variance trade-off.
INTRODUCTION TO MACHINE
LEARNING
WHAT IS MACHINE
LEARNING?
WHAT IS MACHINE LEARNING?
Machine Learning is a subfield of Artificial Intelligence.
ML is based on the model of brain-cell interaction.
The term Machine Learning is coined by Arthur
Samuel, a pioneer in the field of Artificial Intelligence
and computer gaming.
Machine Learning is a field of study that gives
computer “the ability to learn without being explicitly
programmed.
AI vs ML v DL
Artificial Intelligence Machine learning (ML) is a Deep learning (DL) is a
encompasses any specific technique used to subfield of machine learning
computer program achieve AI. inspired by the structure and
that exhibits human- Machine learning algorithms function of the human brain.
like intelligence, allow computers to learn without Deep learning algorithms use
including learning, being explicitly programmed. artificial neural networks with
reasoning, problem- They can improve their multiple layers to process
solving, and decision- performance on a specific task data and extract complex
making. over time by analyzing data. patterns.
AI aspires to create For instance, a recommendation DL is adopted at handling
machines that can system on a shopping website large amounts of data, such
perform tasks might use machine learning to as images, text, or speech.
requiring human analyze your purchase history Facial recognition software is
intellect. and suggest items you'd be an example of a deep
interested in. learning application.
TRADITIONAL PROGRAMMING VS MACHINE LEARNING
MACHINE LEARNING
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 in T, as measured
by P, improves with experience E.
Well posed learning problem
An agent solves a problem or tasks T, Performance P and gain
some experience E
If P is measured at T it can improve Experience E (Learning by
Experience)
Examples:
1. Handwritten Recognition Problem
2. Robo driving learning problem
Problem Task (T) Performance (P) Experience (E)
HANDWRITING
Classifying the A database of
RECOGNITION Better Classification
images and text homework text
LEARNING
Linear Algebra
Linear Algebra
Linear algebra is one of the important branches of
mathematics.
Linear algebra is basically the study of vector spaces,
lines, planes, and linear combinations.
It includes vectors, matrices and linear functions.
Linear algebra helps in manipulating multivariate data.
MULTIVARIATE DATA – INVOLVES MORE THAN TWO
VARIABLES RESULTING IN A SINGLE OUTCOME.
Vector
Vector Spaces
Operations on Vectors
Given two vectors a and b, we form their
sum a+b (Vector Addition)
Commutative of addition: a + b = b + a
Associativity of addition: a + (b + c) = (a + b) +
c
Additive identity: a + 0 = 0 + a = a
Matrix
A matrix is a vector
of vectors
Elements of a
matrix are vectors
m x n matrix have
‘m’ rows and ‘n’
columns
Eigen Value & Eigen Vector
Eigenvalues & Eigenvectors
Eigenvalues are the special set of scalars associated
with the system of linear equations.
The eigenvector is a vector that is associated with a
set of linear equations.
Find Eigenvalue and eigenvector of the matrix,
A = [5 4
1 2]
Spectral Decomposition
Spectral Decomposition
This method decomposes a square matrix ‘A’ into a product of
three matrices.
A = P D P^T
P is a n x n matrix, whose elements are eigenvectors of a
matrix A in terms of decreasing values.
D is a diagonal matrix, whose elements are corresponding
eigenvalues.
Find the spectral decomposition of a matrix,
A = [-3 5
4 -2]
Singular Value Decomposition
