You are on page 1of 10

Introduc)on

to Machine
Learning: Syllabus
Rao Vemuri
rvemuri@gmail.com
Fall 2013

Pre-requisites

Calculus
Chain rule of dieren1a1on. Finding maxima and
minima of func1ons, logarithms, exponen1als

Probability
Sample spaces. Axioms of probability,
condi1onal probability, Bayes formula
Normal distribu1on

MATLAB or equivalent
2

Topic 1. POSING A MACHINE


LEARNING PROBLEM

What is machine learning?


How do you set up the problem?

Topic 2. DECISION TREE LEARNING



Decision tree representa)on, What types of
problems are suitable for decision trees?
Induc)ve bias in decision trees, Examples
You will write a Matlab program to generate a
simple decision tree
You can do a class project using this method

Topic 3. ARTIFICIAL NEURAL


NETWORKS

What are neural nets?


Problems are suitable for neural nets
Perceptrons
Training a Perceptron
Gradient Descent and Delta rule
Mul)-layer networks and Backpropaga)on
Illustra)ve examples
You can do a project using this method
5

Topic 4. EVALUATING A HYPOTHESIS


Es)ma)ng the accuracy of a hypothesis


Basics of sampling theory
Errors in es)ma)on
Bias and variance
Condence intervals
The Central Limit Theorem
Hypothesis tes)ng.
Only ideas. No homework or project on this topic
6

Topic 5. BAYESIAN LEARNING



Bayes Theorem
Maximum Likelihood and Least Squares
Hypotheses
Nave Bayes classier
Learning to Classify Web Documents
This is a good topic for a term project
Minimum Descrip)on Length principle.
Bayes Op)mal Classier
7

Topic 6. GENETIC ALGORITHMS



Gene)c operators, Fitness func)ons and
Selec)on, muta)on, Examples
Schema Theorem
Gene)c Programming
You will do a homework problem on this topic

Text Books
Tom Mitchell, Machine Learning, McGraw Hill,
Second Edi)on
Any good book on MATLAB
I will get you at least two copies of the text
book that students can use as a reference
You can go to Tom Mitchells web site and
download ppt slides and view videos
hdp://www.cs.cmu.edu/~tom/10701_sp11/
lectures.shtml
9

References
There are a number of useful reference books. Each covers material
that is more advanced than the class material.

Bishop, Christopher. Neural Networks for Pa/ern Recogni5on. New
York, NY: Oxford University Press, 1995. ISBN: 9780198538646.
Duda, Richard, Peter Hart, and David Stork. Pa/ern Classica5on.
2nd ed. New York, NY: Wiley-Interscience, 2000. ISBN:
9780471056690.
Has)e, T., R. Tibshirani, and J. H. Friedman. The Elements of
Sta5s5cal Learning: Data Mining, Inference and Predic5on. New
York, NY: Springer, 2001. ISBN: 9780387952840.
MacKay, David. Informa5on Theory, Inference, and Learning
Algorithms. Cambridge, UK: Cambridge University Press, 2003.
ISBN: 9780521642989.

You might also like