Spring 2007 - Machine Learning Reading Group

The machine learning reading group is an informal weekly meeting of those interested in advanced topics in machine learning. We read and discuss a series of recent papers from the ML and related literatures, chosen by group members as interest warrants. The purpose of the meeting is to discuss new ideas, improve all of our understanding, and, hopefully, to have some fun!

Current topics that interest us include, but are not limited to:

Advanced methods for graphical models
- Variational and mean-field methods
- Structure search
- Advanced representations (e.g., relational learning)
Relational, hybrid statistical/symbolic, and other higher-order representation and inference methods
Reinforcement learning and stochastic planning
- Model factoring and decomposition
- Scaling
- Determinism and near-determinism
- MDP and POMDP approximation methods
- Interplay with representation/inference issues
Kernel methods and support vector machines
Non-stationarity, online learning, concept drift, and temporal models
Topological methods
Applications
- Bioinformatics, functional genomics, etc.
- Imaging, vision, handwriting recognition, etc.
- Security, intrusion detection, virus analysis, etc.
- Robotics, control, path plannning, etc.

Schedule and Attendance

The meetings are informal and attendance is completely voluntary -- if you're interested, you're welcome to attend sporadically as your schedule/interest allows. All we ask is that you attempt to read the selected paper for the week before coming to the meeting. Don't feel that you need to fully understand a paper --- a substantial goal is for us all to improve our understanding of these ideas --- but do please try to read the paper in whatever level of detail you're comfortable with.

Mail List

If you would like to receive notifications of papers that we'll be reading, schedule changes, etc. then you should probably sign up for the MLRG mail list.

Meeting time and location

Spring 2007: Fri, 2:30 - 4:00 PM, Farris Engr, Terran's conference room --3rd floor by rooms 323-330b.

This semester's Papers (Spring 2007)

Jan 26: Lin Xiao, Jun Sun, Stephen Boyd "A Duality View of Spectral Methods for Dimensionality Reduction"
Feb 16: G. Elidan and N. Friedman "Learning Hidden Variable Networks: The Information Bottleneck Approach"
Feb 23: (continuing) G. Elidan and N. Friedman "Learning Hidden Variable Networks: The Information Bottleneck Approach"
Mar 2: Jiang Su, Harry Zhang "Full Bayesian Network Classifiers"
Mar 9: (canceled) Mauro Maggioni, Sridhar Mahadevan "Fast Direct Policy Evaluation using Multiscale Analysis of Markov Diffusion Processes"
Mar 23: (continuing) Mauro Maggioni, Sridhar Mahadevan "Fast Direct Policy Evaluation using Multiscale Analysis of Markov Diffusion Processes"
May 4: Mauro Maggioni and Sridhar Mahadevan "Multiscale Framework for Markov Decision Processes using Diffusion Wavelets"

Current candidate papers for this semester

Shohei Shimizu, Patrik O. Hoyer, Aapo Hyvarinen, Antti Kerminen "A Linear Non-Gaussian Acyclic Model for Causal Discovery"

Nir Friedman and Daphne Koller "Being Bayesian About Network Structure. A Bayesian Approach to Structure Discovery in Bayesian Networks."

Chen Yanover, Talya Meltzer, Yair Weiss "Linear Programming Relaxations and Belief Propagation -- An Empirical Study"

Guido Sanguinetti and Neil D. Lawrence "Missing Data in Kernel PCA"

Pascal Poupart, Nikos Vlassis, Jesse Hoey, Kevin Regan "An Analytic Solution to Discrete Bayesian Reinforcement Learning"

Alexander Strehl, Lihong Li, Eric Wiewiora, John Langford, Michael Littman "PAC Model-free Reinforcement Learning"

Rouhollah Rahmani, Sally Goldman "MISSL: Multiple-Instance Semi-Supervised Learning"

Charles Kemp, Joshua B. Tenenbaum, Thomas L. Griffiths, Takeshi Yamada, and Naonori Ueda "Learning Systems of Concepts with an Infinite Relational Model"

Abbeel, Quigley, Ng "Using Inaccurate Models in Reinforcement Learning"

Bowling, McCracken, James, Neufeld, Wilkinson "Learning Predictive State Representations using Non-Blind Policies"

da Silva, Basso, Bazzan, Engel "Dealing with Non-Stationary Environments using Context Detection"

Epshteyn, DeJong "Qualitative Reinforcement Learning"

Bhat, Isbell, Mateas "On the Difficulty of Modular Reinforcement Learning for Real- World Partial Programming"

Kemp, Tenenbaum, Griffiths, Yamada, Ueda "Learning systems of concepts with an infinite relational model"

Guange Dai and Dit-Yan Yeung "Tensor Embedding Methods"

Francois Denis francois, Christophe Nicolas Magnan, Liva Ralaivola "Efficient Learning of Naive Bayes Classifiers under Class-Conditional Classification Noise"

Barbara E. Engelhardt, Michael I. Jordan, Steven E. Brenner "A Graphical Model for Predicting Protein Molecular Function"

Maria-Florina Balcan, Alina Beygelzimer, John Langford "Agnostic Active Learning"

Nima Asgharbeygi, David Stracuzzi, Pat Langley "Relational Temporal Difference Learning"

Ricardo Silva, Richard Scheines "Bayesian Learning of Measurement and Structural Models"

Guange Dai and Dit-Yan Yeung "Tensor Embedding Methods "

Abbeel, Daphne Koller, Andrew Y. Ng "Learning Factor Graphs in Polynomial Time and Sample Complexity Pieter"