CS 8751: Advanced Machine Learning

Lecturer: Rich Maclin (rmaclin)

Syllabus

Grades

Class References:

• Mitchell Textbook Web Page
• Witten and Frank Textbook Web Page
• UCI Repository of ML Data Sets

Class Materials:

• Machine Learning and Knowledge Discovery in Databases - Introduction (Mitchell Chapter 1, Witten Frank Chapter 1)
• Concept Learning: Classification
  • Basics and Version Spaces (M2, WF2, WF3)
  • Decision Trees (M3, WF3.2, WF4.3, WF6.1)
  • Neural Networks (M4)
  • Support Vector Machines   Read Burges' Tutorial (WF6.3)
• Some Statistics: Hypothesis Evaluation (M5, WF5)
• Agent Learning: Reinforcement Learning (M13)
• Unsupervised Learning: Clustering   Clustering Review Paper Read sections 1-5.7, 5.11-5.12 (WF3.9, WF6.6)
• Learning Theory
  • Bayesian Methods (M6)
  • PAC Theory (M7)
• Concept Learning: Classification (Continued)
  • Instance Based Learning (M8, WF3.8, WF4.7, WF6.4)
  • Genetic Algorithms (M9)
  • Rule Learning (M10)
  • Ensemble Learning   Read Opitz & Maclin paper (WF7.4)
  • Mining Association Rules   Read the Apriori Paper

Program Assignments:

1. Program 1: Create Your Own Dataset and Dataset Class
2. Program 2: Learning a Decision Tree for Discrete Features
3. Program 3: Decision Trees - Adding Capabilities to Program 2
4. Program 4: Unsupervised Learning - K-Means Clustering
5. Program 5: Reinforcement Learning - Q Learning for a simple game
6. Extra Credit Program: Genetic Algorithms - A GABIL like learner

Homework Assignments:

1. Homework 1
2. Homework 2
3. Homework 3
4. Homework 4

Sample Exam Questions:

• Sample Exam 1 Questions
• Sample Exam 2 Questions
• Sample Final Exam Questions

Talk Stuff:

• A short list of suggestions on preparing your talk.
• The talk feedback form.
• Guidelines for submitting materials.

Talk Schedule:

Papers to Present:

1. Support Vector Clustering, Ben-Hur, Horn, Siegelmann, and Vapnik, JMLR, 2, 2001, Citeseer PDF
2. A discriminative framework for detecting remote protein homologies, Jaakkola, Diekhans, and Haussler, TR, 1999, Citeseer PDF
3. A Kernel Between Sets of Vectors, Kondor and Jebara, ICML 2003, PDF
4. Fast Training of Support Vector Machines using Sequential Minimal Optimization, Platt, in Advances in Kernel Methods - Support Vector Learning, 1998, Citeseer PDF (NOTE: this should be done as a joint presentation by two students)
5. A Support Vector Machine Formulation to PCA Analysis and its Kernel Version, Suykens, Van Gestel, Vandewalle, and De Moor, TR, 2002, Citeseer PDF
6. Bias Variance Analysis and Ensembles of SVM, Valentini, and Dietterich, Third International Workshop on Multiple Classifier Systems, 2002, Citeseer PDF
7. Inferring a Semantic Representation of Text via Cross-Language Correlation Analysis, Vinokourov, Shawe-Taylor, and Cristianini, NIPS, 2002, Citeseer PDF

1. The following paper is a survey paper discussing several useful ideas, I would like to see a 5 person presentation of this paper with the appropriate supporting papers shown below for each section.

   The paper: Recent Advances in Hierarchical Reinforcement Learning, Barto and Mahadevan, Discrete Event Systems 13, 2003, Citeseer PDF

Topics:

1. The main paper
2. Options

   Based on: Between MDPs and Semi-MDPs: A Framework for Temporal Abstraction in Reinforcement Learning, Sutton, Precup, and Singh, AIJ, 1999, Citeseer PDF

3. Hierarchies of Abstract Machines

   Based on: Reinforcement Learning with Hierarchies of Machines, Parr, and Russell, NIPS, 1998, Citeseer PDF

4. MAXQ

   Based on: Hierarchical reinforcement learning with the MAXQ value function decomposition, Dietterich, JAIR, 2000, Citeseer PDF

5. Multi-agent approaches

   Based on: Hierarchical Multi-Agent Reinforcement Learning, Makar, Mahadevan, and Ghavamzadeh, Fifth International Conference on Autonomous Agents, 2001, Citeseer PDF

2. Model-based Policy Gradient Reinforcement Learning, Wang, and Dietterich, ICML 2003, PDF
3. Bayes meets Bellman: The Gaussian Process Approach to Temporal Difference Learning, Engel, Manoor, and Meir, ICML 2003, PDF

Some Useful Links:

• UCI Repository of ML Data Sets
• OFAI bibliography server (AI reference lookup)
• A somewhat dated set of Links to ML and CBR people
• Oral Presentation Advice