BIRLA INSTITUTE OF TECHNOLOGY AND SCIENCE, PILANI INSTRUCTION DIVISION SECOND SEMESTER 2012-2013 Course Handout Part II Date:

06/01/2013 In addition to part-I (General Handout for all courses appended to the time table) this portion gives further specific details regarding the course. Course No. : BITS C464 Course Title : Machine Learning Instructor- in-charge : NAVNEET GOYAL (goel@bits-pilani.ac.in) Catalog Description Machine Learning is an exciting sub-area of Artificial Intelligence which deals with designing machine which can learn and improve their performance from examples/experience. This course introduces the student to the key algorithms and theory that forms the core of machine learning. The course will cover the major approaches to learning namely, supervised, unsupervised, and reinforcement leaning. The course emphasizes various techniques, which have become feasible with increased computational power. The topics covered in the course include regression, decision trees, support vector machines, artificial neural networks, Bayesian techniques, Hidden Markov models, genetic algorithms etc. Text Books: Tom M. Mitchell, Machine Learning, The McGraw-Hill Companies, Inc. International Edition 1997. Reference Books:
1. Christopher M. Bhisop, Pattern Recognition & Machine Learning, Springer, 2006. 2. Introduction to Machine Learning, N. J. Nilson, Stanford, Available online at authors website. http://robotics.stanford.edu/people/nilsson/mlbook.html 3. Machine Learning, Neural and Statistical Classification, D. Michie, D.J. Spiegelhalter, C.C. Taylor (eds), Ellis Horwood publishers, available online at http://www.amsta.leeds.ac.uk/~charles/statlog/

LECTURE PLAN Topic Overview Preliminaries Topic Details Introduction Probability theory Decision theory Information theory MAP Hypothesis Minimum Description Length (MDL) principle Lecture # 1-2 3-5 Chapter Refere nce Ch. 1 R1 Ch.2

Some important principles/concepts/algori thms

6-8

Ch.6 + class notes

Linear models for Regression Linear models for Classification

Bayesian Learning Techniques Non-linear Models & Model Selection

Expectation Maximization (EM) Algorithm Bias-variance decomposition Lagrange Multipliers Mixture of Gaussians PCA & SVD Linear basis function models Bayesian linear regression Discriminant Functions Probabilistic Generative Classifiers Probabilistic Discriminative Classifiers Bayes optimal classifier Gibbs Algorithm Nave Bayes Classifier Decision Trees Ensemble Classifiers Neural Networks o Multilayer Perceptron o Network training o Error backpropagation Instance-based Learning o K-NN o Case-based Reasoning

9-11 12-15

R1 Ch. 3 R1 Ch. 4

16-17

Ch. 6

18-24

Ch. 3 Ch. 4 R1 Ch. 5 Ch. 8

Margin/Kernel Based Approaches Graphical Models Unsupervised Learning Genetic Algorithms

Class Notes + Support Vector Machines Bayesian Networks Hidden Markov Models Mixture Models K-means Clustering Hypothesis space search 25-27 28-31 32-33 R1 Ch. 7 Ch. 6 + Class Notes Ch. 6 R1 Ch. 9 Ch. 9

34-35

Genetic programming Models of evaluation & learning Q Learning Non-deterministic rewards & actions Temporal difference learning Generalization

Reinforcement Learning

36-37

Ch. 13

Advanced Topics

Active Learning Deep Learning

38-40

Class Notes

Evaluation Scheme: Component Assignments (02) Midsem Test (CB) Comprehensive Exam Duration Take Home 90 Mins. 3 Hours Weightage 35% 25% 40% Date (Time) TBA 25/02 (4.00-5.30 PM) 02/05 (AN)

Notices: All notices shall be displayed on the IPC notice board. Chamber Consultation Hour: T, Th 10 Makeup Policy: To be granted only in case of serious illness. Instructor-in-charge BITS C464