Link Search Menu Expand Document

CS 7650 - Natural Language Processing (Georgia Tech)

Georgia Institute of Technology

  • Instructor: Wei Xu
  • Lecture: Mondays, Wednesdays 3:30-4:45pm

Schedule (subject to change as the term progresses)

Aug 22
Course Introduction
Eisenstein Chapter 1
Aug 24
Machine Learning - naive Bayes [lecture note]
Eisenstein 2.0-2.5, 4.1,4.3-4.5, J+M 5
Aug 25
Problem Set 0 due
Aug 29
Machine Learning - logistic regression
Eisenstein 2.0-2.5, 4.1,4.3-4.5, J+M 5
Aug 31
Machine Learning - Perceptron, SVM, multiclass [lecture note]
Eisenstein 2.0-2.5, 4.1,4.3-4.5, J+M 5
Sep 2
Project 0 due
Sep 5
no class - labor day
Sep 7
Neural Networks - Feedforward [lecture note]
Eisenstein 2.6, 3.1-3.3, J+M 7, Goldberg 1-4
Sep 12
Neural Networks - Backpropagation, optimization
Eisenstein 2.6, 3.1-3.3, J+M 7, Goldberg 1-4
Resource
PyTorch Tutorial (w/ links to Colab notebook)
Sep 13
Problem Set 1 due
Sep 14
Word Embeddings, Course Project
Eisenstein 3.3.4, 14.5, 14.6, J+M 6, Goldberg 5
Sep 19
Sequence Models - HMM, Viterbi [lecture note]
Eisenstein 7.0-7.4, J+M 8
Sep 21
Conditional Random Fields
Eisenstein 7.5, 8.3
Sep 26
Conditional Random Fields (cont’) [lecture note]
Eisenstein 7.5, 8.3
Sep 27
Project 1 due
Sep 28
Recurrent Neural Networks
Eisenstein 7.6, Goldberg 10-11, J+M 9.2, 9.3
Oct 7
Problem Set 2 due
Oct 12
Course Project Proposal due
TBA
[Convolutional Neural Networks + Neural CRFs]
Eisenstein 3.4, 7.6, Goldberg 9
TBA
[Machine Translation]
Eisenstein 18.1, 18.2
TBA
[Encoder-Decoder + Attention]
Eisenstein 18.3 - 18.5
TBA
[Copy/Pointer Network + Transformer]
Vaswani+17 Transformers, Alammar’s blog post, Rush’s tutorial
TBA
[Transformer (cont’) + Course Project]
Holtzman+19 Degeneration
TBA
[Information Extraction]
Eisenstein Chapters 13 & 17
TBA
[Pretrained Language Models]
ELMo BERT
TBA
[Pretrained Language Models (cont’) + Ethics]
BART, T5
TBA
[Dialogue]
J+M Chapter 24, Wizard of Wikipedia
TBA
[Question Answering]
J+M Chapter 23
TBA
[Multilingual / Cross-lingual Models]
XLM-RoBERTa
Dec 9
Final Project Reports Due (submit on Gradescope - no late days allowed)