Natural Language Processing: Instructor: Dr. Muhammad Asfand-E-Yar

Download as pdf or txt
Download as pdf or txt
You are on page 1of 41

Natural Language

Processing
Lecture 1
Instructor: Dr. Muhammad Asfand-e-yar

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


Goals of this subject
• Understanding the challenges of NLP

• Learn about NLP problems and their solution

• Learn general purpose methods

• Be able to read modern research in NLP

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


Disclaimer
I am not inventing the wheel in this class

You are welcome to watch videos, read those if you feel more
comfortable with that
We will change things based on experience, so please be patient.

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


Course Outlines
• Introduction + representing words (word vectors)
• Language models, feed-forward networks, back-propagation
• Tagging, HMMs, Viterbi
• Log-linear models for tagging
• PCFGs, CKY
• Log-linear models for parsing, Lexicalized parsing, dependency
parsing
• Globally normalized log-linear models (structured perceptron, CRFs)
• Semantic parsing
MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar
Recommended books
Speech and Language Processing: An Introduction to Natural Language
Processing: By James H. Martin and Daniel Jurafsky

Ontology Based Interpretation of Natural Language: By Cimiano, Christina


and John

Reference Book:

Statistical Foundations of NLP: By Schutze and Manning:

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


Enter the Selected Term:
Piazza Platform “Spring 2018”

Enter the Access 
Enter the Class code  Code:
as  “ANLPcsc741”
“CSC 741: Adv.NLP”

Press the 
button:
“Join Classes”

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


What is NLP?
Natural-language processing (NLP) is a field of computer science, artificial
intelligence concerned with the interactions between computers and human
(natural) languages, and, in particular, concerned with programming
computers to fruitfully process large natural language data.
Goal: Develop methods for processing, analysing and understanding the
structure and meaning of language.

Structure

Language NLP
Meaning

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


What is NLP?
Application: Build systems that help people do stuff (with text):

Question answering, virtual assistants, translation, etc.

There is a lot language out there…

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


Applications of NLP
1. Search
2. Advertisement matching
3. Translation
4. Sentiment analysis
5. Speech recognition
6. Chatbots
7. Virtual assistants
8. …

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


Levels of Analysis in NLP
Phonology: sounds that make up language

Morphology: internal structure of words

Syntax: structure of phrases, how words modify one another


“pretty little girl’s school”

Semantics: meaning of language in the world


“students sitting on the back row”

Discourse: relations between clauses and sentences


I registered to the class on NLP because it is fascinating
MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar
What’s special about Languages?
Invented by humans for communication

Learned from experience (!!)

A symbolic/discrete system:
table:

piano:

that pre-dates logic…

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


Language is special
Encoded via continuous signals:
• Sounds

• Gestures

• Image (writing)

Brain encodings are continuous

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


Why it is hard?
1. Ambiguity

2. Variability

3. Sparsity

4. Grounding

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


1. Ambiguity
“Finally, a computer that understands you like your mother” (Ad , 1985)

1. The computer understands you as well as your mother


understands you.
2. The computer understands that you like your mother.
3. The computer understands you as well as it understands your
mother.

“Finally, a computer that understands your lie cured mother”


MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar
a) Lexical Ambiguity
Definition:
A word belongs to two or more word (“part of speech”) classes.

Example:
1. the round table (is an adjective),
2. to round the corner (is a verb),
3. dance in a round (is a noun),
4. come round and see us (is an adverb),
5. he walked round the room (is a preposition).

Finite state grammars can be used for resolving lexical ambiguity

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


b) Syntactic Ambiguity
Definition
concern the roles performed by words in sentences and possible
grammatical constructions

Example,
“They are cooking apples” is ambiguous because it may or may
not mean that apples are being cooked …

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


VP

c) Structural Ambiguity
VP V NP

V NP PP discuss violence
PP

discuss violence P NP P NP

on TV
Definition: on TV
A context-free grammar can assign two or more phrase structures (“parse
trees”) to one and the same sequence of terminal symbols (words or word
classes)
Even short sentences have hundreds of analyses
MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar
d) Semantic Ambiguity
Definition:
Concerns the meaning of a word or phrase/sentence
Claims suffer from semantic ambiguity when they contain a word (or words) with
multiple meanings.

Example:
“I don't like it when my father smokes.”
The word “smokes” has more than one meaning, and the significance of the
sentence changes dramatically depending on which meaning is intended.
(smokes also means without results)

“There’s a river that crosses every town.”


Quantification ambiguity (not syntactic)
MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar
1. Ambiguity
Headlines:
• Enraged Cow Injures Farmer with Axe
• Teacher Strikes Idle Kids
• Hospitals Are Sued by 7 Foot Doctors
• Stolen Painting Found by Tree
• Kids Make Nutritious Snacks
more

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


2. Variability
Crucial in semantics
• Dow ends up 255 points

• Dow climbs 255

• All major stock markets surged

• Dow gains 255 points

• Stock market hits a high record

• The Dow Jones Industrial Average close up 255


MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar
3. Sparsity
Definition:
Scattered or Distributed
In NLP we consider the Data Sparsity

Example:
“I ate an apple”
Unigrams: “I”, “ate”, “an”, “apple”
Bigrams: “I ate”, “ate an”, “an apple”

How the given example can be processed?


i.e. “Unigrams” or “Bigrams”

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


Knowledge Base

4. Grounding Relation: has part


Cat Tail
Humans do not learn language by Dog Leg
observation an endless stream of text.
Relation: type of
Ground Words are Knowledge based Tiger Cat
Leg Limb

Relation: instance of
Bengal Tiger
tiger

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


Where NLP lies?
Linguistics

Machine Learning/
Computer Science NLP Statistics

Artificial
Intelligence

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


Related Fields
Computational Linguistics
Use computational models to learn about language (e.g., meaning
change)

Cognitive Science
How does the human brain process language?

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


NLP Applications
1. Text categorization
2. Information Extraction
3. Search
4. Question Answering
5. Virtual assistants
6. Machine translation
7. Summarization
8. Reading comprehension

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


1. Text Categorization

Sports

Politics

Science
MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar
2. Information Extraction
New York Times Co. named Russell T. Lewis, 45, president and general
manager of its flagship New York Times newspaper, responsible for all
business-side activities. He was executive vice president and deputy general
manager. He succeeds Lance R. Primis, who in September was named
president and chief operating officer of the parent.

Person Company Post State


Russell T. Lewis New York Times President and start
newspaper general manager
Russell T. Lewis New York Times Executive vice end
newspaper president
Lance R. Primis New York Times President and end
newspaper CEO

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


3. Search

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


3. Search

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


4. Question and Answering
IBM Watson;
Watson is a question answering computer
system capable of answering questions
posed in natural language, developed
in IBM's DeepQA project by a research
team led by principal investigator David
Ferrucci. Watson was named after IBM's
first CEO, industrialist Thomas J. Watson.
The computer system was specifically
developed to answer questions on
the quiz show Jeopardy! and, in 2011, the
Watson computer system competed
on Jeopardy! against former winners Brad
Rutter and Ken Jennings winning the first
place prize of $1 million.
https://www.techrepublic.com/article/ibm-watson-the-inside-story-of-how-the-jeopardy-winning-supercomputer-was-born-and-what-it-wants-to-do-next/
https://www.youtube.com/watch?v=P18EdAKuC1U
MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar
Critique

Douglas Hofstadter: “just a text search algorithm connected to a


database, just like Google search. It doesn’t understand what it’s
reading.”

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


5. Virtual Assistants

Move all my
Wednesday
meetings in
April

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


6. Machine Translation

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


7. Summarization

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


8. Reading Comprehension
"The rock was still wet. The animal was glistening, like it was still swimming,"
recalls Hou Xianguang. Hou discovered the unusual fossil while surveying rocks as
a paleontology graduate student in 1984, near the Chinese town of Chengjiang.
"My teachers always talked about the Burgess Shale animals. It looked like one of
them. My hands began to shake." Hou had indeed found a Naraoia like those from
Canada. However, Hou's animal was 15 million years older than its Canadian
relatives.
It can be inferred that Hou Xianguang's "hands began to shake“ because he was
a) afraid that he might lose the fossil
b) worried about the implications of his finding
c) concerned that he might not get credit for his work
d) uncertain about the authenticity of the fossil
e) excited about the magnitude of his discovery
MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar
Some History
A. Colorless green ideas sleep furiously
B. Furiously sleep ideas green colorless
“It is fair to assume that neither sentence (1) nor (2) … had ever occurred in an English
discourse. Hence, in any statistical model for grammaticalness, these sentences will be
ruled out on identical grounds as equally “remote” from English. Yet (1), though nonsensical,
is grammatical, while (2) is not.”
S (Chomsky, Syntactic structures,1957)
NP VP

NP

A A N V Adv.

Colorless green ideas sleep furiously


MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar
History; Rule-based
70s and 80s:
• Grammars (rules) of English syntax
• Small domains
• Substantial engineering effort

S → NP | VP N → man
NP → DET | N N → woman
VP → Vt | N DET → the
VP → Vi DET → a
Vt → hugged
Vi → slept
MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar
Empirical Revolution
“Whenever I fire a linguist, our system performance improves.”
Jelinek, 1988

When I look at an article in Russian, I say: “This is really written in English, but it
has been coded in some strange symbols. I will now proceed to decode.’
Weaver, 1995

Of course, we must not go overboard and mistakenly conclude that the successes
of statistical NLP render linguistics irrelevant (rash statements to this effect have
been made in the past, e.g., the notorious remark, “Every time I fire a linguist, my
performance goes up”). The information and insight that linguists, psychologists,
and others have gathered about language is invaluable in creating high
performance broad domain language understanding systems; for instance, in the
speech recognition setting described above, better understanding of language
structure can lead to better language models.”
Lilian Lee, 2001
MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar
Empirical Revolution

1990: corpus-based statistical methods

2000s: rich structural formalisms

2010s: representation learning with neural nets

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


Final Research Work
Groups of exactly two students (MS-Students), one student
(Ph.D. Students)

Options:
• Choose a research paper, implement it
• Build a model that does something you care about
• Define a research problem and work on it
• Must submit a research paper of your interest, related to NLP (that
needs to be approved) on 14th Feb, 2018.

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar


Q&A
That’s all for today’s Lecture

MS(CS), Bahria University, Islamabad Instructor: Dr. Muhammad Asfand-e-yar

You might also like