Introduction to Human Language Technology Philipp Koehn 1 September - - PowerPoint PPT Presentation

introduction to human language technology
SMART_READER_LITE
LIVE PREVIEW

Introduction to Human Language Technology Philipp Koehn 1 September - - PowerPoint PPT Presentation

Introduction to Human Language Technology Philipp Koehn 1 September 2020 Philipp Koehn Introduction to Human Language Technology: Introduction 1 September 2020 Administrative 1 Coordinator: Philipp Koehn (phi@jhu.edu) Lecturers:


slide-1
SLIDE 1

Introduction to Human Language Technology

Philipp Koehn 1 September 2020

Philipp Koehn Introduction to Human Language Technology: Introduction 1 September 2020

slide-2
SLIDE 2

1

Administrative

  • Coordinator: Philipp Koehn (phi@jhu.edu)
  • Lecturers: Faculty of the Center for Language and Speech Processing (CLSP)
  • TA: Desh Raj (r.desh26@gmail.com )
  • Class: Monday, Wednesday, 9:00-10:15pm, MS Teams
  • Course web site: https://jhu-intro-hlt.github.io/
  • Grading

– 5 assignments (10% each) – first midterm exam (15%) – second midterm exam (15%) – final exam (20%)

Philipp Koehn Introduction to Human Language Technology: Introduction 1 September 2020

slide-3
SLIDE 3

2

Course Overview

  • Human Language Technology

– Speech: spoken language (audio) – Text: written language (text)

  • Means of Communication

→ new ways of interacting with computers

  • Storage medium for knowledge

→ new ways of making word knowledge available

  • This course

– methods and tools used in HLT – overview of HLT applications

Philipp Koehn Introduction to Human Language Technology: Introduction 1 September 2020

slide-4
SLIDE 4

3

Course Overview: Text

  • Words, Morphology, Syntax (Yarowsky)
  • Morphology (Yarowsky)
  • Semantics (Post)
  • Deep Learning (Watanabe)
  • Information retrieval and extraction (Koehn, Duh)
  • Machine translation (Duh)

Philipp Koehn Introduction to Human Language Technology: Introduction 1 September 2020

slide-5
SLIDE 5

4

Course Overview: Speech

  • Audio signals, phonemes, graphemes, dictionaries (Elhilali)
  • Auditory system (TBD)
  • Signal processing (Khudanpur)
  • Speech recognition: HMM (Khudanpur)
  • End-to-end neural speech recognition (Watanabe)
  • Speaker identification, language identification (Dehak)

Philipp Koehn Introduction to Human Language Technology: Introduction 1 September 2020

slide-6
SLIDE 6

5

Course Overview: Applications

  • NLP for Digital Humanities (Lippincott)
  • Question answering (Duh)
  • Dialog systems (Sedoc)
  • Clinical NLP (Dredze)
  • Ethical problems (Moro-Velazquez)
  • Analyzing and Interpreting Neural Networks for NLP (TBD)

Philipp Koehn Introduction to Human Language Technology: Introduction 1 September 2020

slide-7
SLIDE 7

6

Master Concentration in HLT

https://www.clsp.jhu.edu/human-language-technology-masters/

  • New this year: Concentration in Human Language Technology

– Master in Computer Science – Master in Electrical and Computer Engineering

  • Requirements (in addition to usual degree requirements)

– Introduction to Human Language Technology (601.667) – Natural Language Processing (601.665) – Information Extraction from Speech and Text (520.666) – Master project in HLT

Philipp Koehn Introduction to Human Language Technology: Introduction 1 September 2020

slide-8
SLIDE 8

7

Center for Language and Speech Processing

  • One of the largest and most influential academic research centers in HLT
  • Faculty in Computer Science, Electrical and Computer Engineering, Cognitive

Science, Mathematical Sciences, ...

  • Home of over 60 researchers, dozens of PhD students
  • Founded in 1992 by Frederick Jelinek (1932-2010)
  • Sibling center: Human Language Technology Center of Excellence (HLTCOE)

Philipp Koehn Introduction to Human Language Technology: Introduction 1 September 2020

slide-9
SLIDE 9

8

Speech Recognition

Philipp Koehn Introduction to Human Language Technology: Introduction 1 September 2020

slide-10
SLIDE 10

9

Information Retrieval

Philipp Koehn Introduction to Human Language Technology: Introduction 1 September 2020

slide-11
SLIDE 11

10

Information Extraction

Philipp Koehn Introduction to Human Language Technology: Introduction 1 September 2020

slide-12
SLIDE 12

11

Machine Translation

Philipp Koehn Introduction to Human Language Technology: Introduction 1 September 2020

slide-13
SLIDE 13

12

Question Answering

Philipp Koehn Introduction to Human Language Technology: Introduction 1 September 2020

slide-14
SLIDE 14

13

Dialog Systems

Philipp Koehn Introduction to Human Language Technology: Introduction 1 September 2020

slide-15
SLIDE 15

14

Hate Speech Detection

incitement of violence / dehumanizing individuals or groups of people

Philipp Koehn Introduction to Human Language Technology: Introduction 1 September 2020

slide-16
SLIDE 16

15

Fake News Detection

Philipp Koehn Introduction to Human Language Technology: Introduction 1 September 2020

slide-17
SLIDE 17

16

Common Themes

  • Hard problems → not solved, but good enough technology
  • Common methods with other subfields of artificial intelligence
  • Technology is advancing rapidly
  • New applications on (and just behind) horizon

Philipp Koehn Introduction to Human Language Technology: Introduction 1 September 2020