Computer Science 248/448

Statistical Speech and Language Processing

Fall 2021

Instructor: Dan Gildea (Office hours: Wed 2:00-3:00pm, Th 2:00-3:00pm, Wegman 3019)
TA: Adam Dees (Office hours: Mon 3:45-4:45pm, Tu 2:20-3:20pm, Wegmans 2307)
Time: Tu/Th 11:05am-12:20pm, 312 Gavett

Homeworks

Recommended resources:

Syllabus

Onwe will coverwhich means that after class you will understandif before class you have read
8/26 Perceptron for Tagging NLP = ML + DP Collins 2002
8/31 Penn Treebank right node raising
9/2 CRFs f - E[f] McCallum 2001
9/7 Hidden Markov Models viterbi, forward JM App A
9/9 Pytorch tensors tutorial
9/14 Neural CRFs convexity Peng 2008
9/16 Context Free Grammars highest prob string, infinite trees JM 13
9/21 latent CRF
9/23 EM for HMM JM App A
9/28 Context Free Grammars viterbi, inside-outside
9/30 Context Free Grammars viterbi, inside-outside
10/5 Parsing minimum risk decoding, max-margin Finkel 08
10/7 Parsing crf parsing, neural parsing Durrett 2015
10/14 Weighted Deduction swiss german Goodman 1999
10/19 Review
10/21 Midterm
10/26 Midterm Solutions
10/28 Search dijkstra, knuth, a*
11/2 Semantics lambda calculus JM 20, Zettlemoyer 2005, Liang 2011
11/4 Transformers attention Vaswani 2017
11/9 Machine Translation BPE, BLEU Papineni 2002, Sennrich 2016
11/11 Machine Translation Model 1 Brown 1993
11/16 Encoder Pretraining bert and bart Bert
11/18 Neural Machine Translation REINFORCE Ranzato 2016
11/23 Semantics word2vec Mikolov 2013, Pennington 2014
11/30 Grammar induction Compound PCFGs Kim, Dyer, and Rush 2019
12/2 Something fun debiasing Rudinger et al. 2018, Zhao et al. 2018
12/7 Review
Final exam: Th 12/16 4-7pm in 101 Hylan.

Grading

  • Homeworks: 40%
  • Midterm: 25%
  • Final: 35%
  • No late homework
  • Programming assignments must be in python, and must run on class account

gildea @ cs rochester edu
December 2, 2021