---

The Unsupervised Learning Group (ULG)

---

What ?
The Unsupervised Learning Group (ULG) is a group of graduate students from the Computer Science and Electrical & Computer Engineering departments, who share interests in data mining, machine learning, information retrieval, pattern recognition and related areas.

When, Where ?
Fall 2003: We meet every alternate week in ACES 2.404B at 5:00 p.m. on Thursdays.

---

Next Meeting:
The next meeting will be at 5pm on Thu, Nov 20, 2003 in ACES 2.404B. The topic of discussion is:

MCMC --- (Misha,Joseph,Razvan)

1. Introduction to Monte Carlo Methods - D.J.C. Mackay

---

The ULG Wish List
The following is a list of topics we want to know more about. Members are requested to volunteer to guide the rest of the group on topics they are comfortable with. [The name(s) in () show the ULGies who voted for a certain topic. The bold voters have volunteered to lead the corresponding discussion.]

• Computational Game Theory --- (Arindam,Srujana,David,Kunal)
  References :
  1. Introduction to Game Theory - Marcus M. Mobius
  2. Computational Game Theory - Michael Kearns
  
  
• PAC Bounds --- (Arindam,Srujana)
  References :
  1. Tutorial on Practical Prediction Theory for Classification - John Langford
  2. PAC-MDL bounds - John Langford

Please send your suggestions/comments to Arindam and/or Sugato.

---

The following is a list of discussions we had and some useful references:

• Reinforcement Learning --- David,Nick
  References:
  1. Reinforcement Learning: A Survey - Leslie Pack Kaelbling, Michael L. Littman, Andrew W. Moore
  2. Towards Learning in Probabilistic Action Selection: Markov Systems and Markov Decision Processes - Manuela Veloso
  3. Reinforcement Learning and Plan Recognition - Manuela Veloso
  
  
• Online Learning --- Srujana
  References:
  1. On-Line Learning - Methods and Open problems - Manfred Warmuth
  2. On-Line Learning and Prediction - Avrim Blum
  3. On-line Algorithms in Machine Learning - Avrim Blum
  
  
• Graphical Models --- Sugato,Joseph
  References :
  1. An Introduction to Graphical Models - Kevin Murphy
  2. The Junction Tree Algorithms - Mark Paskin
  3. Junction Trees - Dennis Bahler
  4. A short course on graphical models - Mark Paskin
  5. Exact inference by Junction/Join/Clique tress - David Page
  
  
• Latent Dirichlet Allocation - Misha
  References :
  1. Latent Dirichlet Allocation - D. M. Blei, A. Y. Ng, M. Jordan
  2. Latent Dirichlet Allocation - D. M. Blei, A. Y. Ng, M. Jordan
  
  
• Variational Methods for Graphical Models - Shi
  References :
  1. An Introduction to Variational Methods for Graphical Models - M. I. Jordan, Z. Ghahramani, T. S. Jaakkola, L. K. Saul
  
  
• Boosting - Arindam
  References :
  1. A decision-theoretic generalization of on-line learning and an application to boosting - Yoav Freund, Robert Schapire
  2. Boosting the margin: A new explanation for the effectiveness of voting methods - Robert Schapire, Yoav Freund, Peter Bartlett, Wee Sun Lee
  
  
• Non-negative Matrix Factorization
  References :
  1. Algorithms for Non-negative Matrix Factorization - Daniel Lee, Sebastian Seung
  2. Learning the parts of objects by non-negative matrix factorization - Daniel Lee, Sebastian Seung
  3. Computational neuroscience: Think positive to find parts - B. W. Mel
  
  
• Product of Experts - Jefferson
  Refernces :
  1. Training Products of Experts by Minimizing Contrastive Divergence - Geoffrey E. Hinton
  2. Recognizing Hand-written Digits Using Hierarchical products of Experts - Guy Mayraz, Geoffrey Hinton
  3. Rate-coded Restricted Boltzmann Machines for Face Recognition - Yee Whye Teh, Geoffrey Hinton
  
  
• Co-training - Sugato
  References :
  1. Combining Labeled and Unlabeled Data with Co-Training - Avrim Blum, Tom Mitchell
  2. Analyzing the Effectiveness and Applicability of Co-training - Kamal Nigam, Rayid Ghani
  
  
• Object Discovery through Motion - Joseph
  No references since the work is still unpublished.
  
  
• Random projections and Mixture of Gaussians - Mitul
  References :
  1. Experiments with Random Projection - Sanjoy Dasgupta
  2. Learning Probability Distributions - Sanjoy Dasgupta, PhD Thesis
  
  
• Kamal Nigam's thesis - Shi
  Thesis: Using Unlabeled Data to Improve Text Classification - Kamal Nigam
  
  
• Clustering
  Paper : Clustering based on conditional distributions in an auxiliary space - Janne Sinkkonen and Samuel Kaski
  
  
• The Online Median Problem - Ramgopal
  References :
  1. The Online Median Problem - R. R. Mettu, C. G. Plaxton
  2. Optimal Time Bounds for Approximate Clustering - R. R. Mettu, C. G. Plaxton
  
  
• VC Dimensions - Rupert
  References :
  1. Probably Approximately Correct Learning - Haussler
  2. Bounds on the Sample Complexity of Bayesian Learning Using Information Theory and the VC Dimension - Haussler, Kearns, Schapire
  3. Learnability and the Vapnik-Chervonenkis Dimension - A. Blumer, A. Ehrenfeucht, D. Haussler, and M.K. Warmuth, Journal of ACM, 36(4): 929-965, 1989.
  
  
• XML - Anuj, Srujana
  References :
  1. Language and the Internet - Rajesh Bhatt's course page
  2. XML and its standarization efforts
  3. Research directions - Widom
  4. other resources - W3C site, database and products
  
  
• How Naive is the Naive Bayes Classifier - Sugato
  References :
  1. On the Optimality of the Simple Bayesian Classifier under Zero-One Loss - Domingos, Pazzani
  2. A Comparison of Event Models for Naive Bayes Text Classification - McCallum, Nigam
  3. On Bias, Variance, 0/1 - loss, and the Curse-of-Dimensionality - Friedman
  
  
• Expectation Maximization - Misha
  References :
  1. A Gentle Tutorial of the EM algorithm and its Application to Parameter Estimation for Gaussian Mixture and HMM Models - Jeff A. Bilmes
  2. Foundations of Statistical Natural Language Processing - Manning, Schutze: Section 14.2.2
  
  
• Clustering using Graph Partitioning - Alex
  References :
  1. Part I : Basics, BFS, KL - J. Dammel
  2. Part II : Spectral, Multi-level - J. Dammel
  
  
• Suffix Trees - Arindam
  References :
  1. WUM : A Web Utilization Miner - Spiliopoulou, Faulstich

---

The following are the members of the group:

• Sreangsu Acharyya --- (sreangsu(at)ece.utexas.edu)
• Arindam Banerjee --- (abanerje(at)ece.utexas.edu)
• Sugato Basu --- (sugato(at)cs.utexas.edu)
• Patrick Beeson --- (pbeeson(at)cs.utexas.edu)
• Mikhail Bilenko --- (mbilenko(at)cs.utexas.edu)
• Dan Fernholz --- (fernholz(at)cs.utexas.edu)
• Gurushyam Hariharan --- (guru(at)lans.ece.utexas.edu)
• Alex Hennegue --- (hennegue(at)ece.utexas.edu)
• Rohit Kate --- (rjkate(at)cs.utexas.edu)
• Anuj Khare --- (khare(at)ece.utexas.edu)
• Matt MacMahon --- (adastra(at)mail.utexas.edu)
• Prem Melville --- (melville(at)cs.utexas.edu)
• Ramgopal Mettu --- (ramgopal(at)cs.utexas.edu)
• Srujana Merugu --- (merugu(at)ece.utexas.edu)
• Joseph Modayil --- (modayil(at)cs.utexas.edu)
• Aniket Murarka --- (aniket(at)cs.utexas.edu)
• Un Yong Nahm --- (pebronia(at)cs.utexas.edu)
• Krupakar V Pasupultei --- (krupakar(at)cs.utexas.edu)
• Jefferson Provost --- (jp(at)cs.utexas.edu)
• Kunal Punera --- (kunal(at)lans.ece.utexas.edu)
• Arvind Rao --- (ukarvind(at)cs.utexas.edu)
• Suvrit Sra --- (suvrit(at)cs.utexas.edu)
• Siddharth Sriram --- (sid(at)hercules.ece.utexas.edu)
• Alexander Strehl --- (strehl(at)hercules.ece.utexas.edu)
• Rupert Tang --- (rupert(at)cs.utexas.edu)
• Mitul Tiwari --- (mitult(at)cs.utexas.edu)
• Vamsi K Vutukuru --- (vamsikv(at)cs.utexas.edu)
• Yuk Wah Wong --- (ywwong(at)cs.utexas.edu)
• Joseph Xavier --- (xavier(at)ece.utexas.edu)
• Qingfeng Yu --- (yuqf(at)hercules.ece.utexas.edu)
• Jun Yuan --- (jyuan(at)cs.utexas.edu)
• Shi Zhong --- (szhong(at)ece.utexas.edu)

---

page maintained by: Arindam Banerjee