Advanced Topics in Machine Leaerning

CS 546
Advanced Topics in Machine Learning
Spring Quarter 2010


Time : Tuesdays and Thursdays, 2:00-3:50pm
Location: Campus Rec. Center (ASRC), Room 220.

Instructor: Melanie Mitchell, FAB 120-24, (503) 725-2412, mm-AT-cs.pdx.edu.
Office hours: Tuesdays and Thursdays, 4:00-5:00pm, or by appointment.

Course Website: :
http://www.cs.pdx.edu/~mm/aml2010/index.html

Prerequisites: CS 445/545, or equivalent.

Projects: Each student will carry out a term-long research project in some area of machine learning. Each student will give a presentation and write a paper on their project by the end of the term. Each student will also be assigned to review/edit some number of other students' papers.

Paper presentations: Each student will present in class some number of recent papers related to their project

Reading: There will be no textbook. Required reading will handed out in class or accessible via the Web.

Exams: There will be no exams.

Grading: Students will be graded on their class presentations, class participation, and final project talk and paper.

Academic integrity: Students will be responsible for following the PSU Student Conduct Code, and in particular, the policy concerning academic honesty.

Students with disabilities: If you are a student with a disability in need of academic accommodations, you should register with Disability Services for Students and notify the instructor immediately to arrange for support services.

Class schedule (subject to change):

Date

Topics and Reading

Tuesday March 30

Course intro

Discussion of research areas

Bayesian Networks

Reading:
S. Wooldridge, Bayesian Belief Networks

Thursday April 1

Hidden Markov models

Principal Components Analysis

Tuesday April 6

Students present research proposals and reading lists

MM presents on Using Analogy to Discover the Meaning of Images.

Reading:
Serre et al. Robust object recognition with cortex-like mechanisms

Optional reading:
Russell and Norvig, Chapter 15, sections 1-3.

L. I. Smith, A tutorial on Principal Components Analysis

Thursday April 8

Topic: Understanding Classification Decisions of Additive Classifiers
Presenter: Will Landecker
(MM away)

Reading:
Poulin et al, Visual Explanation of Evidence in Additive Classifiers. Conference on Innovative Applications of Artificial Intelligence, July 2006.

Robnik-Sikonja and Kononenko, Explaining Classifications for Individual Instances. IEEE Trans. on Knowledge and Data Engineering, May 2008, vol. 20(5).

Tuesday April 13

Topic: Weighted and Fuzzy SVMs, Semi-supervised SVMs

Presenter: Josh Hoak

Reading:
Lin and Wang, Training algorithms for fuzzy support vector machines with noisy data

Wang and Chiang, Fuzzy support vector machine for multi-class text classification

Thursday April 15

Topic: Simultaneous localization and mapping for embedded agents

Presenter: Spenser Barlow

Reading:
Montmerlo et al., FastSLAM: A factored solution to the simultaneous localization and mapping problem

Dellaert et al. Monte Carlo localization for mobile robots

Optional:Arulampalam et al., A Tutorial On Particle Filters for Online Non-Linear Non-Gaussian Bayesian Tracking

Tuesday April 20

Topic: Evolution and coevolution of cellular automata and random Boolean networks

Presenter: Alireza Goudarzi

Reading:

Mitchell, M (2006). Coevolutionary learning with spatially distributed populations

Pagie, L. and Mitchell, M. (2002). A comparison of evolutionary and coevolutionary search

Carnevali, S. and Patarnello, P. 1989. Learning Networks of Neurons with Boolean Logic

Thursday April 22

Topic: The Baldwin effect:  Interactions between evolution and learning

Presenter: Jason Akers

Reading: Genes, Phenes and the Baldwin Effect: Learning and Evolution in a Simulated Population . RM French and A Messinger. Artificial Life IV 1994.

Lamarckian Evolution, The Baldwin Effect, and Function Optimization. D. Whitley, VS Gordon, K Mathias

Optional: A Hybrid Self-Adaptive Genetic Algorithm Based on Sexual Reproduction and Baldwin Effect for Global Optimization. Zhang, Zhao, Wang. IEEE 2009.

Tuesday April 27

Topic: Genetic algorithms for feature selection

Presenter: Jeff Weston

Reading: Oh et al., Hybrid genetic algortihms for feature selection

Derrac et al., IFS-CoCo: Instance and Feature Selection Based on Cooperative Coevolution with Nearest Neighbor Rule

Thursday April 29

Topic: Extensions of the HMAX model

Presenter: Jason Hall

(MM away)

Reading: J. Mutch and D. G. Lowe, Object class recoginition and localization using sparse features with limited receptive fields

Wu et al., Object recognition by learning informative, biologically inspired visual features

T. Masquelier and S. J. Thorpe, Unsupervised learning of visual features through spiketiming dependent plasticity

Tuesday May 4

Topic: MicroRNAs, cancer, and multi-class classification

Presenter: David Gibbs

Reading:
Freund, An adaptive version of the boost by majority algorithm

Lu et al., MicroRNA expression profiles classify human cancers

Thursday May 6

Topic: Classifying and designing peptides

Presenter: Chris Whelan

Reading:
C. Leslie, E. Eskin, J. Weston, and W. Noble, Mismatch string kernels for SVM protein classification, Advances in Neural Information Processing Systems, pp. 1441–1448, 2003.

L. Jacob, B. Hoffmann, V. Stoven, and J.-P. Vert, Virtual screening of GPCRs: an in silico chemogenomics approach, BMC Bioinformatics 2007, vol. 9, p. 363, Jan 2008.

Tuesday May 11

Topic: Model order and cue combination in visual recognition

Presenter: Karan Sharma

Reading:
A. Rabinovich, A. Vedaldi, and S. Belongie, Does Image Segmentation Improve Object Categorization?

A. Rabinovich, T. Lange, J. Buhmann, and S. Belongie, Model Order Selection and Cue Combination for Image Segmentation

Thursday May 13

Topic: Hierarchical Temporal Memory (HTM)

Presenter: Ryan Price

Reading:
George and Jaros, The HTM learning algorithms

George and Hawkins, Toward a mathematical theory of cortical micro-circuits (optional)

Tuesday May 18

Topic: TBA

Presenters: Max Orhai and Bert Shaw

Reading:

For Bert's presentation

EJ Pauwels and G Frederix, Finding salient regions in images. Nonparametric clustering for image segmentation and grouping

HD Cheng and Y Sun, A hierarchical approach to color image segmentation using homogeneity


For Max's presentation

M. Tomassini, Generalized Automata Networks

S. Harding et al., Self-Modifying Cartesian Genetic Programming

Thursday May 20

Topic: Feedback and attention in the HMAX model

Presenter:  Max Quinn

Readings:

F. Miau and C. Papageorgiou, Neuromorphic algorithms for computer vision and attention

Tuesday May 25

No class

Thursday May 27

First draft of project papers due; papers assigned to reviewers

Topic: System identification with recurrent neural networks

Presenter: Josh Hughes

Readings:

R. Santiago, Context Discerning Multifunction Networks: Reformulating Fixed Weight Neural Networks

N. E. Cotter and P. R. Conwell, Learning Algorithms and Fixed Dynamics

A. S. Younger et al., Meta-Learning with Backpropagation

Tuesday June 1

Topic: Motor control and memory

Presenter: Ryan Mitchell

Readings:

Gomez and Miikkulainen, Solving Non-Markovian Control Tasks with Neuroevolution

Gomez et al., Efficient Non-Linear Control through Neuroevolution

Thursday June 3

Reviews of project papers due

Student project presentations

Monday June 7

10:15-12:05: Student project presentations

Wednesday June 9 (Make-up class)

10:15-12:05: Student project presentations

Note change of location to FAB 150.

Friday June 11

Final project papers due