Download as pdf or txt
Download as pdf or txt
You are on page 1of 15

Natural Language Processing

CS 1462

Introduction
Computer Science
3 rd s e m e s t e r - 1 4 4 4
Dr. Fahman Saeed
faesaeed@imamu.edu.sa

1
Today

 Overview of the field


 Overview of course
 Background knowledge

 Topics

 Activities and Grading

 Administrative Stuff

 Introductions (if time left)

CPSC
2
503 –
Winter
2019
3/20/2023
Assessment

Percentage of Total
# Assessment task* Week Due
Assessment Score
1 Midterm Exam 7th 20%
2 Assignments 2nd : 10th 10%
3 Project 11th 15%
Quizzes' (1 & 2) 5th and 10th 15%
4

5 Final Exam 12th 40%


Total 100%
Textbooks

[SLP2] Daniel Jurafsky and James H. Martin. 2009. Speech and


Language Processing (2nd Edition). Prentice-Hall, Inc., USA. (The 3rd
Required edition is coming out soon and should be adopted instead)
Textbooks
[SLP3] Daniel Jurafsky and James H. Martin. 2009. Speech and
Language Processing (3rd Edition).

Essential
[IANLP] Nizar Y. Habash , Introduction to Arabic Natural Language
References
Processing, Morgan & Claypool, 2007
Materials
Instructor Policies

• Handouts will be uploaded to the course web site (blackboard).


• Any announcement will be posted on blackboard. It is your responsibility
to check it once in a while to keep yourself updated.
• It is necessary to attend the class regularly to build a comprehensive
foundation of the course. If you exceed the limit of absence you will not
be allowed to enter the final exam.
• Second time late class coming; student will be considered absent for 1
day.
• No makeup quizzes and workshops for any reason. If you do not show up
you will get a zero.
• Contact me: an official email - office hours.

5
Instructor Policies

• No makeup midterm and final exams. If you have an approved


circumstance, you will be allowed to enter a comprehensive exam at the
end of this semester. The exam will include the whole curriculum.

• You should present your medical excuses, signed by the committee of


excuses, no later than a week from your absence. Otherwise, it will not
be accepted.

• Studying is from the book; correct answers are drawn from the text
book.

6
Office Hours

Be sure to check the calendar before coming or send email!


Natural Language Processing

 What is it?
We’re going to study formalisms,
models and algorithms to allow
computers to perform useful tasks
involving knowledge about human
languages.

CPSC
8
503 –
Winter
2019
3/20/2023
Sample Useful Tasks

 Any ideas?

CPSC
9
503 –
Winter
2019
3/20/2023
Sample Useful Tasks

 Conversational agents: AT&T “How may I


help you?” technology
 Apple SIRI…..
 Summarization: ”Please summarize my
discussion with Sue about 503” “What people
say about the new Nikon 5000?”
 Yahoo Paid $30 Million in Cash for the Summly
company (2013)
 Generation: an automatic commentator of a
soccer game (e.g., from output of a vision
system)
 ARRIA world leader in NLG- when it floated on London's
Alternative Investment Market (AIM) in 2013, it was valued at
over £160 million

3/20/2023
CPSC 503 – Winter 2019 10
Sample Useful Tasks (cont’)

 Web-based question answering : “Was 1991 an El Nino


year? ….Was it the first one after 1982?” “Why was it so
intense?”
 IBM Watson Jeopardy (now medicine! See next slides)
 Document Classification: spam detection, news filtering
 Information Extraction ……

…Machine Translation

CPSC
11
503 –
Winter
2019
3/20/2023
 Watson : analyzes natural language questions and
content well enough and fast enough to compete and
win against champion players at Jeopardy!

• NLP systems performing useful NLP tasks are very


complicated involving many algorithms (some non-
NLP) and many of datasets
• When someone tells you that they have solved NLP or
even just a subtask of NLP be always skeptical
• Transferring results form a game to useful / general
solutions is extremely difficult

Source:IBM

CPSC
Slid
422,
e 12
Lecture
35
Not just NLP…

• Parsing (PCFGs)
• Shallow parsing (NP segmentation with CRFs)
• Entity and relation Detection (NER with CRFs)
• Logical Form Generation and Matching
• Logical Temporal and Spatial Reasoning
• Leveraging many databases, taxonomies, and
ontologies (help only 25% of questions)
• Confidence… Probabilities (Bnets to rank)
• Strategy for playing Jeopardy…statistical models of
players and games, game-theoretic analyses … .. and
application of reinforcement-learning (Buzz-in - Bets)

CPSC
Slid
422,
e 13
Lecture
35
14
Wishing you a fruitful educati
onal experience

15

You might also like