CS 294-5: Statistical Natural Language Processing, Fall 2005

 
Instructor: Dan Klein
Lecture: Mondays and Wednesdays, 1:00-2:30pm, 310 Soda Hall
Office Hours: Mondays and Wednesdays 2:30-3:30pm in 775 Soda Hall, or by appointment

Announcements

11/4/05: Homework 5
10/18/05: Homework 4
10/18/05: Section on 10/21 in Soda, 1-2pm, on word alignment
10/15/05: Extension: Homework 3 due on Wednesday 10/19

10/4/05:  Homework 3
10/1/05:  Update: No class on 10/3 or 10/5 (HW2 still late if timestamped after 10/3)
9/26/05:  No class on 10/5
9/26/05:  Invite: StatNLP lunch, Tuesdays 12:30 in Soda 373 [topic]
9/26/05:  Final project guidelines
9/19/05:  Homework 2
9/19/05:  Reminder: my office hours are cancelled on Tuesday, but I'll be back on Wednesday.
9/14/05:  Update: Aria's office hours will be extended to F 12-3 in Soda 493, at least this week.
9/12/05:  Aria's office hours will be F 12-1 in Soda 493
9/11/05:  My office hours have moved, by popular demand, to T 11-12, W 2:30-3:30
9/02/05:  Want a Millennium account?  Fill out the form by Tuesday morning if you want it soon.
8/31/05:  Problems with the newsgroup?  Check here.
8/31/05:  Homework 1
8/31/05:  Accounts and access
8/29/05:  Class policies
8/29/05:  Class questionnaire
8/16/05:  The course newsgroup is ucb.class.cs294-5. If you use it, I'll use it!
8/16/05:  The previous website has been archived.

Description

This course will explore current statistical techniques for the automatic analysis of natural (human) language data. The dominant modeling paradigm is corpus-driven supervised learning, but unsupervised methods and even hand-coded rule-based systems will be mentioned when appropriate.

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 estimate parameters for 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 familiarity with basic probability and the ability to program in Java. Prior experience with linguistics or natural languages is helpful, but not required.

Readings

The texts for this course are:

  • Manning and Shuetze, Foundations of Statistical Natural Language Processing [amazon.com] [online version]
  • Jurafsky and Martin, Speech and Language Processing [amazon.com]

The former is loosely required (i.e. you'll want access to a copy) while the latter is recommended as supplementary reading.  Both are on reserve in the Engineering library.

Syllabus

Week Date Topics Techniques Readings Assignments (Out) Assignments (Due)
1 Aug 29 Course Introduction M+S 1-3
Aug 31 Language Models Multinomial Smoothing M+S 6, J+M 6,
Chen & Goodman
HW1: Language Models
2 Sep 5 NO CLASS  
Sep 7 Language Models EM, More Smoothing
3 Sep 12 Text Categorization Naive-Bayes M+S 7, Event Models
Sep 14 Word-Sense Disambiguation Maximum-Entropy Berger's tutorial
4 Sep 19 Text Clustering EM HW2: PNP Classification HW1
Sep 21 Part-of-Speech Tagging HMMs M+S 9-10, J+M 7.1-7.4
5 Sep 26 Part-of-Speech Tagging MEMMs / CRFs Toutanova & Manning,
Brants, Brill
Sep 28 Word Class Induction Distributional Models HMM Learning, Distributional Clustering
6 Oct 3

NO CLASS

HW3: POS Tagging HW2
Oct 5

NO CLASS

7 Oct 10 Machine Translation (Transfer) IBM Models M+S 13, J+M 21, IBM Models.
Oct 12 Machine Translation (Transfer+Decoding) IBM Models HMM, Decoders
8 Oct 17 Machine Translation Phrase-Based Models Phrase-Based
Oct 19 Syntactic Parsing / Ambiguities M+S 3.2, 12.1, J+M 12 HW4: Machine Translation HW3
9 Oct 24 Unlexicalized PCFGs Splitting Methods Unlexicalized, M+S 12.1
Oct 26 Parsing Algorithms CKY M+S 11
10 Oct 31 Lexicalized Parsing M+S 12.2, J+M 12.3-4, Best-First, A*, Collins, Charniak and Johnson
Nov 2 Semantic Interpretation Compositional semantics, J+M 14,15 HW5: Parsing / Grammars HW4
11 Nov 7 Semantic Roles, Coreference Semantic Role Labeling, Empty Reconstruction, Coreference  
Nov 9 Grammar Induction Model Merging, Distributional, Constituency/Dependency, Translingual Constraint
12 Nov 14 Question Answering
Nov 16 Information Extraction HW5
13 Nov 21 Syntactic Translation
Nov 23 The Speech Signal
14 Nov 28 Speech Recognition
Nov 30 Final Projects (NIPS groups)
15 Dec 5 Speech Synthesis    
Dec 7 Final Project Presentations   FP