CS 294-19: Statistical Natural Language Processing, Spring 2008

Instructor: Dan Klein
Lecture: Tuesday and Thursday, 11:00am-12:30pm, 320 Soda Hall
Office Hours: Tuesday and Thursday 12:30-1:30pm in 775 Soda Hall.
GSI: Aria Haghighi
Section: Wed. 12-1pm in 320 Soda
Office Hours: Wed. 1-5pm in 525 Soda Hall (or by appointment)


3/17/08:  Assignment 5 is posted, due April 3.
2/17/08:  Telebears mistakenly kicked out half the class.  I'm working on it -- but don't panic, you're all in the class.
2/16/08:  Assignment 3 is posted, due Feb 28. A few days extension is likely, but start early as your experiments will be compute intensive!
2/12/08:  New section time! Friday 4-5pm in 320 Soda, starting this week.
2/4/08:  Assignment 2 is posted, due Feb 14.
2/3/08:  Section time survey for possible new section time. Please vote even if you like the current time. This week (2/6) section still W 12-1pm.
2/2/08:  Newsgroup exists!
2/2/08:  Readers now available at Copy Central at Hearst and Euclid.
1/29/08:  Section is now Wednesday 12-1pm in 320 Soda.
1/21/08:  Assignment 1 is posted, due Feb 5.
1/21/08:  The course newsgroup is ucb.class.cs294-19. If you use it, I'll use it!
1/21/08:  The previous website has been archived.


This course will explore current statistical techniques for the automatic analysis of natural (human) language data. The dominant modeling paradigm is corpus-driven statistical learning, with a split focus between supervised and unsupervised methods.

In the first part of the course, we will examine the core tasks in natural language processing, including language modeling, word-sense disambiguation, morphological analysis, part-of-speech tagging, syntactic parsing, semantic interpretation, coreference resolution, and discourse analysis. In each case, we will discuss which linguistic features are relevant to the task, how to design efficient models which can accommodate those features, and how to learn with such models in data-sparse contexts. In the second part of the course, we will explore how these core techniques can be applied to user applications such as information extraction, question answering, speech recognition, machine translation, and interactive dialog systems.

Course assignments will highlight several core NLP tasks. For each task, we will construct a basic system, then improve it through a cycle of linguistic error analysis and model redesign. There will also be a final project, which will investigate a single topic or application in greater depth. This course assumes a good familiarity with basic probability and a strong ability to program in Java. Prior experience with linguistics or natural languages is helpful, but not required.  Disclaimer: there will be a lot of statistics and algorithms in this class, as well as some serious coding.


The primary texts for this course are:

I recommend both, but everything that you need is online.

Syllabus [subject to substantial change!]

Week Date Topics Techniques Readings Assignments (Out) Assignments (Due)
1 Jan 22 Course Introduction (6PP) (2PP) J+M 1, M+S 1-3 HW1: Language Models
Jan 24 Language Modeling (6PP) (2PP) Multinomial Smoothing J+M 4, M+S 6, Chen & Goodman  
2 Jan 29 Language Modeling (6PP) (2PP) More Smoothing Interpreting KN  
Jan 31 Text Classification (6PP) (2PP) Naive Bayes M+S 7, Event Models
3 Feb 5 Word Sense Disambiguation (6PP) (2PP) Maximum Entropy Classification Tutorial Maxent Tutorial 1, 2, J+M 6 HW2: PNP Classification HW1
Feb 7 Classification (6PP) (2PP) J+M 5, Toutanova & Manning,
Brants, Brill


4 Feb 12 Part-of-Speech Tagging (6PP) (2PP) HMMs/CRFs J+M 6, M+S 9-10, HMM Learning, Distributional Clustering    
Feb 14 Word Classes (6PP) (2PP) EM Johnson HW3: POS Tagging HW2
5 Feb 19 Speech Recognition (6PP) (2PP) Speech Signal J+M 7  


Feb 21 Speech Recognition (6PP) (2PP) Acoustic Modeling J+M 9    
6 Feb 26 Machine Translation (6PP) (2PP) Word Alignment J+M 25, IBM Models, HMM Agreement Discriminative HW4: Machine Translation
Feb 28 Machine Translation (6PP) (2PP) Word Decoding Decoding HW3
7 Mar 4 Machine Translation (6PP) (2PP) Phrase-Based Systems Decoding, Learning Phrases    
Mar 6 Parsing I (6PP) (2PP) M+S 3.2, 12.1, J+M 11    
8 Mar 11 Parsing II (6PP) (2PP) M+S 11, J+M 12 HW5: Parsing  
Mar 13 PCFGs (6PP) (2PP) Unlexicalized, Split HW4
9 Mar 18 Lexialized Parsing (6PP) (2PP) M+S 12.2, J+M 12.3-4, Best-First, A*, Collins, Charniak and Johnson
Mar 20 Grammar Induction (6PP) (2PP)    
10 Mar 25 Spring Break
Mar 27 Spring Break
11 Apr 1 Semantic Roles J+M 16, 19  


Apr 3 Coreference Supervised, Unsupervised, J+M 21  HW5
12 Apr 8 Compositional Semantics (6PP) (2PP) Manning, J+M 18  FP Guidelines  
Apr 10 Semantic Interpretation Parsing to LF  
13 Apr 15 Question Answering (6PP) (2PP)  
Apr 17 Question Answering
14 Apr 22 Syntactic Translation GHKM, Vs Phrases, Decoding
Apr 24 Grammar Induction
15 Apr 29 Historical Reconstruction / Phrase Learning Sampling    
May 1 Final Presentations  
16 May 6 Final Presentations      
May 8 Conclusion / Translation from Monotexts?     FP