Machine Learning

Class: S.Y. B.Tech.

Course Code: DJ19AMC403

An adaptation of NPTEL lectures and online resources for educational purpose

Books Recommended:

Text books: 1. Tom M.Mitchell ―Machine Learningǁ McGraw Hill,2017.

2. Peter Harrington ―Machine Learning In Actionǁ, DreamTech Press,2012.

3. Ethem Alpaydın, ―Introduction to Machine Learningǁ, MIT Press, 2014.

Reference Books:

1. Han Kamber,―Data Mining Concepts & Techniquesǁ, Morgann Kaufmann Publishers, 2011.

2. 2. Stephen Marsland, ―Machine Learning An Algorithmic Perspectiveǁ CRC Press, 2015

3. 3. Kevin P. Murphy , Machine Learning ― A Probabilistic Perspectiveǁ, 2012.

4. 4. Andreas C. Müller and Sarah Guido- Introduction to Machine Learning with Python: A Guide for Data Scientists, O'reilly,2016
Evaluation Scheme:

Semester End Examination (A):

Theory: 1. Question paper will be based on the entire syllabus summing up to 75 marks.

2. Total duration allotted for writing the paper is 3 hrs.

Continuous Assessment (B): Theory:

1. Two term tests of 25 marks each will be conducted during the semester

2. Laboratory: (Term work) Term work shall consist of minimum 8 experiments.

3. The distribution of marks for term work shall be as follows:

4. i. Laboratory work (Performance of Experiments): 15 Marks

5. ii. Journal documentation (Write-up and/or Assignments): 10 marks The final certification and acceptance of term work will be subject to satisfactory performance of laboratory
work and upon fulfilling minimum passing criteria in the term work
Programmatic Solution Vs Machine Learning
Example: Sum of 2 nos.

Scenarios:

C program:

1.Spam: if else:

2. Image classify: Dog

Intelligence:

Data mining, Analysis: graphs. Hidden info,

Why ML: Data, Computing power, Hardware, Algorithms

AI/ML/DL:
Subjects:
AI/ML/DL:
Intelligence: pattern recognition, IQ, creative, imagination, emotional intelligence, define love?
pattern recognition:
symbolic system: knowledge representation, expert system, ex. chess, inference engine
diagram:
specific case. but not work with fuzzy logic, image, voice
so ML: diff symbolic AI n ML .. no explicit programming ( rule based)but data output n find program
many photos of dog, learning from pattern
no need to write rules but just provide data n apply statistics
DL: same as ML but diff algo. biology inspired algo
neuron, perceptron
ML problem: features provide, example: dog n cat features define but if DL features automatically create & support fuzzy logic which is a problem in ML
new layers create efficiency prediction improve, go more deeper,
Example: resume placement: features like certificate, percentage, backlog, placement person but DL will automatically extract features
digit identify: 4 : edges in 1st layer, then so on.
more data more performance improvement
but ML stops at some point
why ML use example: knife vs talvar
Machine Learning

Popularity of this field in recent time and the reasons behind that
–New software/ algorithms
•Neural networks
•Deep learning
–New hardware
•GPU’s
–Cloud Enabled
–Availability of Big Data
 Unit 1

Introduction to Machine Learning

Machine Learning : Definition

Learning is the ability to improve one's behaviour based on experience

•Build computer systems that automatically improve with experience

•What are the fundamental laws that govern all learning processes?
•Machine Learning explores algorithms that can
–learn from data / build a model from data
–use the model for prediction, decision making or solving some tasks

Learning is used when:

• Human expertise does not exist (navigating on Mars),

• Humans are unable to explain their expertise (speech recognition)
• Solution changes in time (routing on a computer network)
• Solution needs to be adapted to particular cases (user biometrics)
Machine Learning : Definition

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.
[Mitchell]
Components of a learning problem

Task: The behaviour or task being improved.

–For example: classification, acting in an environment

•Data: The experiences that are being used to improve

performance in the task.

•Measure of improvement :
–For example: increasing accuracy in prediction, acquiring
new, improved speed and efficiency
What is Machine Learning?

• Optimize a performance criterion using example data or past

experience.

• Role of Statistics: Inference from a sample

• Role of Computer science: Efficient algorithms to

• Solve the optimization problem

• Representing and evaluating the model for inference

Black-box Learner
Learner
ML Applications
ML Applications
Choosing a Model Representation

• The richer the representation, the more useful it is for

subsequent problem solving.
• The richer the representation, the more difficult it is to
learn.

• Components of Representation
–Features
–Function class / hypothesis language
 Machine Learning History

1950s:
–Samuel's checker-playing program
1960s:
–Neural network: Rosenblatt's perceptron
–Minsky & Papert prove limitations of Perceptron
1970s:
–Symbolic concept induction
–Expert systems and knowledge acquisition bottleneck
–Quinlan’s ID3 (Iterative Dichotomiser 3)
–Natural language processing (symbolic)
 Machine Learning History

1980s:
–Advanced decision tree and rule learning
–Learning and planning and problem solving
–Resurgence of neural network
–Valiant’s PAC learning theory
–Focus on experimental methodology
90's ML and Statistics
–Data Mining
–Adaptive agents and web applications
–Text learning
–Reinforcement learning
–Ensembles
–Bayes Net learning
Machine Learning History

Popularity of this field in recent time and the reasons behind that
–New software/ algorithms
•Neural networks
•Deep learning
–New hardware
•GPU’s
–Cloud Enabled
–Availability of Big Data