Computer Science 248/448
Homework
Fall 2015
 Homework 1, Due Tu 9/8 11am

Implement the perceptron decoder using data and weights given in /u/cs448/data/pos.
What is your accuracy on the test file?
 Homework 2, Due Th 9/10 11am

Implement perceptron training using data given in /u/cs448/data/pos.
What is your accuracy on the test file when training on the train file?
 Homework 3, Due Tu 9/29 11am

Implement CRF training for POS tagging using data from the previous
assignment, and the same feature set.
Use SGD to optimize the weights.
What is your accuracy on the test file when training on the train file?
 Homework 4, Due Tu 10/13 11am

Implement latent variable CRF training for POS tagging using data from the previous
assignment, and the same feature set.
Spilt each tag into (at least) two latent tags.
What is your accuracy on the test file when training on the train file?
 Homework 5, Due Th 10/22 in class

Review problems handed out in class.
 Homework 6, Due Tu 11/10 11am

Implement IBM Model 1, and plot log probability of training
and test files in /u/cs448/data/mt.
Here is a simple test case.
 Homework 7, Due Tu 11/17 11am

Implement Gibbs samplingbased EM for IBM Model 1.
 Homework 7, Due Tu 11/24 11am

Implement the fertility HMM model of Zhao and Gildea (EMNLP 2010).
 Homework 8, Due Th 12/10 in class
 Review problems handed out in class
gildea @ cs rochester edu
December 8, 2015