601.467/667 Introduction to Human Language Technology


Fall 2022

Coordinator: Philipp Koehn (phi@jhu.edu)
TA: Yunmo Chen (yunmo@jhu.edu)
CAs: Rosalyn Shin (hshin36@jhu.edu), Venkat Mukthineni (vmukthi1@jhu.edu)
Class: Tuesday and Thursday 9:00-10:15am, Gilman 132 50
Office hours: Coordinator: on request
Office hours:
Yunmo: Friday, 11:00 am - 12:00 pm, Malone 216 / Zoom
Venkat: Monday, 3:00 pm - 4:00 pm, Malone 216 / Zoom
Rosalyn: Tuesday, 3:30 pm - 4:30 pm, Malone u216 / Zoom
(Zoom-only session would be announced in advance on Piazza)
Gradescope (entry code: DJKDYD)PiazzaLecture recordings

Assignments

Note: For Fall 2022, please *do not* proceed before assignments being announced. They are subject to change.
You can confirm whether the homework is released by checking the date listed on each homework.

Late submissions: For each student, we allow a total of 10 days of late submission for all homeworks.
It is counted on a daily basis, for example, if you submit a homework even a few minutes late, you will lose 1 day of your quota.
After you use up all 10 days of late submission, each late day would cost 5% points penalty.
Late submission for teamwork would use 1-day for each teammate.
For each homework, you *are not allowed to submit* after 20 days.
  1. N-gram language modeling, CYK parsing: Due on September 28 (Wednesday)
  2. RNNLMs, word2vec: Due on October 22 (Saturday)
  3. Seq2seq for pronunciation prediction: Due on November 20 (Sunday)
  4. Speech recognition with CTC: Due on December 14 (Wednesday)

Exam

There will be two mid-terms and final exam. You are allowed to bring 1 sheet of paper with notes to the exam. The final exam time is December 14, 6-7:15pm

Lectures

Date Topic Instructor
Tu Aug 31IntroductionKoehn
Text
Th Sep 1Words and Language ModelsYarowsky
Tu Sep 6MorphologyYarowsky
Th Sep 8SyntaxPost
Tu Sep 13SemanticsLippincott
Th Sep 15Machine TranslationPost
Tu Sep 20Deep learning IMurray
Tu Sep 22Distributional SemanticsKoehn
Th Sep 27Deep learning II (Python notebook)Murray
Th Sep 29Information RetrievalDuh
Tu Oct 4Information ExtractionKoehn
Th Oct 6First Midterm Exam-
Speech
Tu Oct 11Auditory systemElhilali
Th Oct 13Speech basicsMoro-Velazquez
Th Oct 18Speaker recognitionVillalba
Tu Oct 25Enhancement and DiarizationGarcia
Tu Oct 27Classic speech recognition1 (additional slides)Khudanpur
Th Nov 1End-to-end neural speech recognitionKhudanpur
Tu Nov 3Hands on: Kaldi (K2, ESPnet)Khudanpur
Tu Nov 8Hands on: Kaldi (continued)
Th Nov 10Second Midterm Exam
Applications
Tu Nov 15Question AnsweringDuh
Th Nov 17NLP for Digital HumanitiesLippincott
Tu Nov 29(no class) 
Th Dec 1Ethical ProblemsMoro-Velazquez
Tu Dec 6Is Scale all You Need?Daniel Khashabi
Th Dec 8Clinical NLPDredze
1These slides present an incomplete picture of what will be discussed in class. Attentive listening is recommended for gaining maximal benefit.