CMPUT 466/551
Auxiliary Material

The material below is related to various topics covered in CMPUT 466/551, to augment the slides .
In column2:  [B:x] = refer to chapter x of the Bishop textbook;
[A:x] = Alpaydin, [DHS:x]=Duda/Hart/Stork, [M:x]= Mitchell, [HTF:x]= Hastie/Tibshirani/Friedman, [WF:x]=Witten/Frank and [RN:x] = Russell/Norvig.

See also material from other courses

Various Texts Articles / Slides / Course /Book
 Other (Web)

Introduction: What is ML? 
[A:1], [DHS:1], [M:1], [HTF:1,2], [RN:18.1]

Probability 101 [HTF:1,2]
[B: 2; AppendixB], [A:1], [DHS:1], [M:1], [RN:18.1]
Book: Grimstead/Snell: "Introduction to Probability"
On-line text: Concepts and Applications of Inferential Statistics
Slides: (from A. Moore)

STAT 221 (UofA: Kouritz)

PDF: Cribsheet

Web: Conditonal Probably and Bayes' Rule
Relationships between (Clickable) Inventory of Probability Distributions
Univariate Dist'n Rel'ns
Conjugate Priors
Conjugate #2
Learning Linear Regression Models [3,7]
[DHS: 3] [A:?]

General Issues for Predictors [3.2, 3.4]
Domingos, AAAI 2000
Domingos, ICML 2000

Linear Classifiers [4]
[A: 11.2, 5.8, 10.1, 10.3, 105.-10.7; 4.1, 4.2, 4.5; 5.1, 5.2, 5.4-5.6], [HTF:5], [M:4], [RN:19-19.5]
Slides: Linear Classifiers (Moore)
          DHS Ch5 Slides
          Chapter (from T Mitchell)
Framework: Supervised Learning, Issues
Hypothesis Testing for Evaluating Predictors
[A: 14],[DHS:9], [M:5] DHS Ch9 Slides (ppt) Web: Simple Interactive Statistical Analysis
A New View of Statistics
Discussion of ROC curves
Statistical Engineering website
Artificial Neural Nets  
(BackPropagation, Repr'n, Conjugate Gradient, Topics)
[B: 5], [A: 11], [M:4], [DHS:6], [RN:20.5]
Slides: DHS Ch6 Sildes
Article: Intro to Conjugate Gradient w/out ... Pain
Web: Neural Nets FAQ
Decision Trees
(Entropy, Pruning, Topics)
[B: 14.4] [A: 9.1-9.2], [DHS:9], [M:3], [RN:18.3]
Book: C4.5: Programs for Machine Learning
Web: AIxploratorium
Kernel Trick [6]
?Gaussian Process? [A: ?] [HTF ?4.5.2, 12.2-12.3], ...
Web: Gaussian Process Slides: Dan Lizotte
Linear Algebra, etc: Matrix Cookbook
Cheat Sheet
Support Vector Machines
(Quadratic Programming, Kernel Trick, ...)
[HTF: 12]
[B: 6, 7, AppE], [A: 10.9], [HTF 4.5.2, 12.2-12.3]
Slides: SVM (R Meir)
Auxiliary notes
Books: Vapnik: "The nature of statistical learning theory" [12 Xeroxed pages]
     Kernel Methods for Pattern Analysis Support Vector
Articles: Intro to Kernel-Based Learning Algs
Web: Kernel Machines
Freeware: SVM Light
Lagrange Multipliers:
       SLIMY GA Tech Wikipedia
Computational Learning Theory
[A: 2.1-2.3], [M:7], [RN:18.5]

Book:  An Introduction to Computational Learning Theory 
Ref:  Probabilistic inequalities
Course: Blum's "Machine Learning Theory"
Ensemble Methods
(Bagging, Boosting)
[B: 14.3], [A: 15], [DHS: 9.5], [HTF:10] Slides: Ensemble Learning
from Ensemble Methods in Machine Learning (Dietterich)
Notes from S Wang
Web: Shapire, movie
AdaBoost Demo
Boosting Papers [UCSD]
Dietterich's Page
Papers on Boosting/Bagging (Schapire)
Article: "An Introduction to Boosting and Leveraging" (Meir/Raetsch), On Bagging and Nonlinear Estimation (Friedman,Hall)
Directed Models (Belief Nets) [8]
[A: 3.7]  [RN:20.2]
[2.11], [3.10], [M:?], [RN:?]
Slides: Cmput366-Notes on BeliefNets
GMM (Moore)
HMM (Moore)
Articles: Heckerman: A tutorial on learning with Bayesian networks
       BN - Classifiers
      HMM (Rabiner)
Web: Belief Nets
       Interactive Tutorial
       Ten years of HMMs
Undirected Models
(Markov Random Fields)
[8.4] MaxEnt (Berger)
References (Wang)
Unsupervised Learning  
  • K-means clustering
  • Principle Components Analysis
  • Independent Component Analysis
[12], [A: 7], [DHS:10], [HTF:14]

Web: Tutorial
Reinforcement Learning
  • Markov Decision Process, Known Model,
    Adaptive Dynamic Programming,
    Value/Policy Iteration
  • TD-Learning
    Value Approximation
  • Q Learning
  • [16], [M:13], [RN:20]
    Book: Reinforcement Learning: An Introduction (download)
    Slides: Barto slides
    ArticlesKaelbling's survey article
           Misc pointers
           POMDPs for Dummies
    Web:Dietterich's Reinforcement Learning
           UMass Amherst
           CMU Reinforcement Learning Group