MLRG - FALL 2007

Welcome to the Machine Learning Reading Group at Brown University. The meeting time this semester is Thursday, 12-1. We will meet in the CIT third floor atrium.

MLRG meetings and other news are announced via posts to the ML-READING-GROUP list. You can subscribe at The MLRG leader this semester is Micha Elsner.

Our topics this semester are, provisionally, spectral clustering, variational Bayes and modeling with dynamical systems. However, we are continuing the option of 'clinic' sessions in which one member presents a research problem and we discuss various proposed solutions. If you have an ongoing or just-completed project involving machine learning, please offer to lead a session!

Date Topic Important People
Sept. 6 Organizational Meeting
Bring your schedule!
Sept. 13 Spectral Clustering Basics
On Spectral Clustering (read sec. 1,2 and look at fig. 1)
A Tutorial on Spectral Clustering (skim sec. 1,2,3,4)
and consider Spike's problem from the email! (python)
Slides on Graph Diffusion (expanded, with Matlab code and more visuals: PPT) and K-Means
Spike, David Jackson
Sept. 20 Spectral Clustering (Slightly Less) Basics
On Spectral Clustering (continue to read)
Graph Partitioning Notes
Read till you get confused!
Spike, David Jackson
Sept. 27 Learning the Number of Clusters (and More...)
Spectral Methods for Automatic Multiscale Clustering
Eric Sodomka, Ian Sherman
Oct. 3 Learning the Number of Clusters (Comparison with Parametrics)
Continue with Spectral Methods for Automatic Multiscale Clustering
Also Learning the K in K-means
the Anderson-Darling test Wikipedia
No official leader
Oct. 11 Semisupervised Spectral Learning
Read Spectral Learning
Supplementary: Semisupervised Learning using Gaussian Fields and Harmonic Functions
Dan Grollman
Oct. 18 No-regret Learning in Convex Games
No reading required!
For background on no-regret, you could check out No-Phi Regret: A Connection Between Computational Learning Theory and Game Theory
or other readings and slides from Amy's course last spring
Casey Marks
Oct. 25 Adaptive Distance for K-NN
Improving nearest neighbor rule with a simple adaptive distance measure
David Jackson
Nov. 1 Dynamical Models: Intro
Traffic Flow in 1D Cellular Automaton Model Including Cars Moving with High Speed
Watch the pretty cars: simulations using the Intelligent Driver Model
Slides (PDF)
Micha Elsner, Jason Pacheco
Nov. 8 Dynamical Models: Language Change
The Logical Problem of Language Change
Micha Elsner (other volunteers welcome)
Nov. 15 Dynamical Models: Navigation
Behavioral Dynamics of Steering, Obstacle Avoidance and Route Selection
William Warren
Nov. 29 Kalman Filtering
Maybeck: Stochastic Models, Estimation and Control, chap. 1 (read first)
Parameter Estimation for Linear Dynamical Systems (if you have time)
Micha Elsner (I guess)
Dec. 6 More Kalman Filtering
matlab code (look at optionally)
Jason Pacheco

We're not going to get to all our suggested papers.

Paper ideas: