CS322 (Natural Language Processing) Spring 2012, Carleton College

Week 1: introduction to NLP

10 September 2012 (M): introduction to NLP and to the course.
Reading: Jurafsky–Martin, Chapter 1.

PS01 (background survey and summarizing a document) was posted on Moodle today. It's due Wednesday at 12:01am.
12 September 2012 (W): whirlwind tour of linguistics; ps01 debrief.
Reading: Jurafsky–Martin, Chapter 2. Concentrate on §2.1; read §2.2–2.3 as a preview.
Viewing: Four Candles (The Two Ronnies).
14 September 2012 (F): regular expressions (see http://regexpal.com/), tokenization, stemming, and Eliza.
PS02 (pig latin and gibberish) was posted on Moodle today. It's due Wednesday before class, with your partner (assigned via email today).

Reading: Jurafsky–Martin, §3.1–3.9, especially §3.1, §3.8, and §3.9.
For the original Porter stemmer paper (and some implementations), see Martin Porter's page about his stemmer.

Week 2: word distributions, edit distance, and language models

17 September 2012 (M): Zipfian distributions, spelling corrections, edit distance.
Reading: Jurafsky–Martin, §3.9–3.10.
http://en.wikipedia.org/wiki/Wikipedia:AutoWikiBrowser/Typos
19 September 2012 (W): edit distance and extensions, intro language modeling.
Reading: Jurafsky–Martin, §4.0–4.3.

PS03 (spelling correction) was posted to Moodle today. It's due ~~Wednesday (or earlier)~~ next Friday.
21 September 2012 (F): probability crash course, language modeling, n-grams.
Reading: nothing new, but you can check out more on Dasher.
Come to class Monday with a proposal of how to evaluate a language model.

Week 3: language models, smoothing

24 September 2012 (M): n-grams (http://books.google.com/ngrams/datasets); computing n-grams; smoothing.
Reading: Jurafsky–Martin, §4.4–4.9; optionally finish Chapter 4.
26 September 2012 (W): smoothing n-grams.
Reading: Jurafsky–Martin, §5.0–5.3.
28 September 2012 (F): ps04, evaluating language models (perplexity), spelling correction in context.
Reading: Three article relevant to ps04 (read in as much detail as you find interesting):
- Jack Hitt, "Words on Trial". The New Yorker, 23 July 2012. (See Moodle.)
- Jasmine Novak, Prabhakar Raghavan, Andrew Tomkins. Anti-Aliasing on the Web. WWW'04.
- Myle Ott, Claire Cardie, Jeff Hancock. Estimating the Prevalence of Deception in Online Review Communities. WWW'12.
PS04 (identifying authorship) was posted on Moodle today. It's due next Monday (October 8th) before class, with your partner (assigned via email today). However, the concepts in the problem set are fair game for the exam on Friday, so I strongly encourage you to work on it before then.

Week 4: part-of-speech tagging; Hidden Markov Models

1 October 2012 (M): part-of-speech tagging and Hidden Markov Models.
Reading: Jurafsky–Martin, §6.1–6.4; §5.5

The first exam is on Friday, in class.
3 October 2012 (W): HMMs: executing and decoding; the forward algorithm and the backward algorithm.
Reading: Jurafsky–Martin, §6.5 (for Monday).
5 October 2012 (F): prelim #1

Week 5: HMMs

8 October 2012 (M): HMMs: Viterbi algorithm; thinking about training.
Reading: review Jurafsky–Martin, §6.0–6.5.
10 October 2012 (W): training HMMs: forward–backward algorithm/Baum–Welch.
Reading: Jurafsky–Martin, §12.0–12.4.

Resubmit ps04 following the correct interface from class today.

PS05 (POS tagging) was posted on Moodle today. It's due next Friday, October 19th, before class, with your partner (assigned via email today).
12 October 2012 (F): ps04 debrief; midterm course evals; context-free grammars begun.
Reading: Jurafsky–Martin, §13.0–13.3, §14.1; skim rest of Chapter 13.

Week 6: syntax and the beginning of sound

15 October 2012 (M): midterm break!
17 October 2012 (W): ps05 questions; context-free grammars and SCIgen; the barest hints of sound processing.
Reading: Jurafsky–Martin, Chapter 7.

Extension granted on question #3 of PS05 (POS tagging) until ~~Monday~~ Wednesday. Submit your report.pdf with the second part of your assignment.
19 October 2012 (F): phonetics and phonology overview; sound overview; whistling into microphones.
Reading: Jurafsky–Martin, §9.0–9.1

Further extension granted on question #3 of PS05 (POS tagging) until Wednesday.

Week 7: speech processing

22 October 2012 (M): ps05 chat; speech reminder; sound processing begun.
Reading: Jurafsky–Martin, §9.3.0–9.3.3; wikipedia on DFT

Go to Sarah Keller's Phi Beta Kappa talk (Using Basic Science to Understand 21st Century Energy Options) on Tuesday (7:00p, Olin 149).

Go to Loren Terveen's CS collquium talk (Crowdsourcing: From Analysis to Design to Evaluation) on Thursday (4:00p, CMC 206).
24 October 2012 (W): architecture of speech recognition systems; DFTs begun.
PS06 (writing papers automatically) was posted on Moodle today. It's due Monday before class, with your partner (assigned via email today).

Reading: Jurafsky–Martin, §9.3.4.
26 October 2012 (F): Class cancelled!

Week 8: speech processing

29 October 2012 (M): DFTs.
Reading: Jurafsky–Martin, §9.3.5–9.3.6; §9.4–9.6.
31 October 2012 (W): DFTs and dialtones.
Reading: Nothing new.

PS07 (speech recognition) was posted to Moodle today. It's due ~~next Monday~~ Wednesday 11/7.
2 November 2012 (F): FFTs.
Reading: wikipedia on FFTs.

Week 9: speech processing

5 November 2012 (M): FFT and speech processing wrapups.
Reading: skim Jurafsky–Martin, Chapter 10 (especially §10.3). Read Jurafsky–Martin, Chapter 8.
7 November 2012 (W): automatic speech recognition wrapup; take-home exam; final project.
Prelim #2 (the take-home exam) was distributed today. It's due on Monday.

The final project announcement was posted on Moodle today. Your topic proposal is due by Sunday at noon.
9 November 2012 (F): Class replaced by office hours!

Week 10: speech synthesis; wrapup

12 November 2012 (M): speech synthesis systems.
The final project groups/topics were announced by email today. Your group will present during our finals slot (Monday, 19 November, 3:30pm–6:00pm).
14 November 2012 (W): final presentation guidelines; course evaluations; ask me anything.