This schedule is subject to change.

Assignments are due right before class (12:30pm) on their due date unless noted otherwise.

Readings can be done either before or after class (unless noted otherwise); they are chosen to support the material covered in class.

Week 1: An introduction

• Aug 29: Introduction, and a demo of some computational linguistics tasks and methods.
• The demo is here.

Week 2: Using rules to analyze individual words

• Sep 3: Introduction to programming: Python basics
  • Worksheet: first steps in Python.
• Reading: Think Python chapter 1 and chapter 2
  Warning: This online book uses Python 2, not Python 3. The main difference that you will see is that it uses print() without the (). This doesn't work in Python 3!
  
  
• Sep 5: Morphology, finite state automata and finite state transducers
• Taking words apart: What do "sing", "singing", "sings", "sang" have in common, and how do we describe what is different?
• Reading: J&M ch. 2 second half, ch. 3, pp. 45-68

Week 3

• Sep 10: Introduction to programming: Writing functions in Python
• Worksheet: Building your own Python functions
• Readings: Think Python chapter 3, chapter 5

• Sep 12: Finite state automata, finite state transducers, and morphology
  
• Assignment 1 due
  

Week 4: Rules again, this time to build dialog systems

• Sep 17: Introduction to programming: Python conditions and lists
  
• Worksheet: Python conditions, lists, and loops
  
• Reading: Think Python chapter 4

• Sep 19: Regular expressions, and using text patterns in dialog systems
• Python worksheet: regular expressions
• Jurafsky and Martin 3rd edition chapter 2, section 2.1
  
• Dialog systems:
• Eliza and Parry examples, another Eliza example
  
• Implementing Eliza

Week 5: Machine learning

• Sep 24: Introduction to programming: Counting words with Python dictionaries and with the Natural Language Toolkit
• Worksheet: Python dictionaries
• Using Python dictionaries: mapping strings to lists
• The nltk.FreqDist data type is described in chapter 1 of the NLTK book.

• Sep 26: Machine learning. Method: Naive Bayes classifiers. Application: Sentiment analysis.
  
• Demo: conditional probability
• Reading: J&M edition 3, chapter 4 (available online by this link)
• Assignment 2 due

Week 6: Word sequences

• Oct 1: Introduction to programming: Loose ends: file handling, and sorting. Also, list comprehensions (which you don't have to use, but which are pretty)
• Worksheet: Reading and writing files in Python
• Worksheet: Sorting in Python
• Python list comprehensions
• Reading: Think Python chapter 6 and chapter 12

• Oct 3: Machine learning continued.
• In class: course project discussion. Bring your ideas!
• Assignment 3 due

Week 7: Word meaning

• Oct 8: Typical word sequences: Language models
• Methods: Probabilistic approaches, n-grams
  Reading: J&M ch. 4, pp 83-97
  Demo: Language models in Python

• Oct 10: Automatically determining word meaning: word embeddings
• Reading: J&m 3rd edition chapter 6, available from this link.
• Slides.
  
• Assignment 4 due

Week 8: Parts of speech

• Oct 15: Word embeddings continued, and a second look at language models
• Some ready-made distributional models you can use for your project, if needed: scroll down to "Where to get a pretrained model"
• The general introduction to neural networks was based on this chapter of the book "Neural networks and deep learning"
• Slides (pages 31 and following)
  
  
• Oct 17: Is this a noun or a verb? Part-of-speech tagging.
• Readings: NLTK book chapter 5 section 4; 
  J&M ch. 2 first half
• Assignment 5 due
  

Week 9

• Oct 22: Sequence information in Part-of-speech tagging
  
• Methods: Hidden Markov Models (HMMs), dynamic programming
  
• Reading: J&M ch. 5 pp. 139-149, ch. 6 pp. 173-179, NLTK book ch. 5 sec. 5.5
  
• Hidden Markov Models for POS-tagging in Python
  

• Oct 24: Training HMMs
• Method: Expectation Maximization
• Demo: the forward-backward algorithm
• Intermediate project report due
  

Week 10: Syntactic structure

• Oct 29: Chunking and parsing
  
• Reading: NLTK book ch. 7, sec. 7.2 and 7.3
• Reading: J&M ch.12, pp. 385-404

• Oct 31: Describing the syntax of English with context-free grammars
• Assignment 6 due
  
  

Week 11

• Nov 5: Automatic syntactic analysis: Parsing
• Reading: J&M ch. 13, pp. 427-450, NLTK book ch. 8
  
  
• Nov 7: Automatic syntactic analysis: Parsing
• Assignment 7 due

Week 12

• Nov 12:  Statistical parsing
• Method: Probabilistic context-free grammar
  
• Reading: J&M ch. 14 pp. 459-463, 467
• Here is a phrase structure parser that you can try out. It uses the LinGO English Resource Grammar.
  If you would like to see a sentence with PP attachment issues, try: "We quickly put up the tents in the lee of a small hillside and cook for the first time in the open.
• Here is a dependency parser that you can try out. It uses Stanford CoreNLP, a tool that you can also download.
• We also discussed Universal Dependencies, a project that develops cross-linguistically unified dependency annotations for many different languages.
  

• Nov 14:  Guest lecture: Ray Mooney on RoboNLP
  
• Assignment 8 due
  
  
• Nov 15 11:00 Computational Linguistics colloquium talk by Yoav Artzi
  

Week 13: Meaning and structure

• Nov 19: The Chomsky hierarchy, or: how complex is human language?
• Reading: J&M, ch. 16, pp 529-542
  

• Nov 21: Meaning and structure: semantic roles, and semantic parsing
• Resources:
  
• PropBank semantic roles in the Unified Verb Index
• FrameNet semantic roles
• Semantic proto-roles at the Decompositional Semantics Initiative
• Abstract Meaning Representations (AMR)
  
• A tutorial on semantic parsing
  
  
• Nov 22 11:00 Computational Linguistics colloquium talk by Jacob Andreas
  

Week 14: Applications

• Nov 26L Applications: Machine Translation
• Reading: J&M ch. 25, p. 859-876
• Material:  Slides: Chris Brew on neural machine translation, Chris Brew on neural machine translation with attention
  
• Final project report due