Introduction#

What is NLP?#

  • Building computer systems that understand and generate natural languages.

  • Deep understanding of broad language

    • not just string processing or keyword matching

Can you think of NLP Applications?

Speech Recognition#

Speech Recognition is usually not considered NLP. We will not cover this topic here.

beach

recognise speech vs. wreck a nice beach

Sentiment Analysis#

sent

Machine Translation#

mt

http://translate.google.com/

Information Extraction#

ie1

Information Extraction#

ie2

Generation#

gen

Generation#

gen

Summarisation#

../_images/summarization2.png

Question Answering#

../_images/qa.png ../_images/qa2.png

Reading Comprehension#

../_images/comprehension.png

Personal Assistants#

Siri1 Siri2

What is difficult about NLP?#

Discuss and enter your answers here: https://ucph.page.link/nlp_q1

(Responses)

Ambiguity Everywhere#

  • Fed raises interest rates 0.5% in effort to control inflation

  • Fed raises interest rates 0.5% in effort to control inflation

  • Fed raises interest rates 0.5% in effort to control inflation

Ambiguity Everywhere#

“Jane ate spaghetti with a silver spoon.”

Do you mean…

Jane used a silver spoon to eat spaghetti? (cutlery)

Jane had spaghetti and a silver spoon? (part)

Jane exhibited a silver spoon while eating spaghetti? (manner)

Jane ate spaghetti in the presence of a silver spoon? (company)

Ambiguity on different linguistic levels#

../_images/nlp_pyramid.png

Core NLP Tasks#

  • Tokenisation, Segmentation

  • Part of Speech Tagging

  • Language Modelling

  • Machine Translation

  • Syntactic and Semantic Parsing

  • Document Classification

  • Information Extraction

  • Question Answering

Core NLP Methods#

  • Structured Prediction

  • Preprocessing

  • Generative Learning

  • Discriminative Learning

  • Weak Supervision

  • Representation and Deep Learning