Machine Learning for NLP

Wintersemester 2011/12
Thomas Müller, Hinrich Schütze
Th 14:00-15:30, M12.21

Schedule

date topic questions literature
Oct 20 Overview, Basics of Probability Theory, the Expectation-Maximization (EM) Algorithm pdf Paskin. Introduction to Probability Theory . 2003.
Oct 27 Exponential Families, Maximum Entropy (ME), Kullback-Leibler divergence pdf Sudderths. Graphical Models for Visual Object Recognition and Tracking. 2006. (2.1 Exponential Families)
Nov 3 Baysian Inference, Conjugate Priors, Multinomial/Dirichlet-Distribution Sudderths. Graphical Models for Visual Object Recognition and Tracking. 2006. (2.1 Exponential Families) ME example
Nov 10 Multinomial/Dirichlet-Distribution Sudderths. Graphical Models for Visual Object Recognition and Tracking. 2006. (2.1 Exponential Families)
Nov 17 Unigram Model, Multinomail Mixture Model, P-LSI pdf Hoffmann. Probabilistic Latent Semantic Indexing. 1999.
Nov 24 De Finetti, LDA Blei, Ng, Jordan. Latent Dirichlet Allocation. 2003.
Dec 1 LDA, Variational Inference Blei, Ng, Jordan. Latent Dirichlet Allocation. 2003.
Dec 8 Markov Chain Monte Carlo (MCMC), Gibbs-Sampling pdf Gilks, Richardson, Spiegelhalter. Markov Chain Monte Carlo in Practice. 1996. (1 Introducing MCMC)
Dec 15 LDA, Gibbs-Sampling Griffiths, Steyvers. Finding scientific topics. 2004.
Dec 22 Practical exercise: LDA, PageRank, ME lda. External LDA implementations: lda-c and gibbslda
Jan 12 Decision trees and coreference resolution I
Jan 19 Unsupervised Part-of-Speech (PoS) Tagging, Gibbs-Sampling Goldwater, Griffiths. A Fully Bayesian Approach to Unsupervised PoS Tagging. 2007.
Jan 26 Decision trees and coreference resolution II
Large Margin Methods Noah Smith, Linguistic Structure Prediction, Section 3.6
Feb 2 Sentiment analysis
main paper: Modeling online review
background: Multigrain models
Feb 9 Maximum entropy, Stanford MaxEnt classifier slides
0
1
2
3