View File - UET Taxila

Mobile and Pervasive
Computing - 7
Natural Language
Processing
Presented by: Dr. Adeel Akram
University of Engineering and Technology, Taxila,Pakistan
http://web.uettaxila.edu.pk/CMS/AUT2016/teMPCms
Outline

Natural Language Processing

Human Computer Dialog Systems

Problems and Success in HCD

Machine Translation

Example based Machine Translation

Projects
What is Natural Language
Processing?

NLP is an interdisciplinary field that uses computational
methods to:

Investigate the properties of written human language and
model the cognitive mechanisms underlying the understanding
and production of written language.

Develop novel practical applications involving the intelligent
processing of written human language by computer.
What is NLP? (cont.)


NLP plays a big part in Machine learning techniques:

automating the construction and adaptation of machine
dictionaries

modeling human agents

essential component of NLP

closer to AI
We will focus on two main types of NLP:

Human-Computer Dialogue Systems

Machine Translation
Human-Computer Dialogue
Systems

Usually with the computer modeling a human dialogue
participant

Will be able:

To converse in similar linguistic style

Discuss the topic

Hopefully teach
Current Capabilities of
Dialogue Systems


Simple voice communication with machines

Personal computers

Interactive answering machines

Voice dialing of mobile telephones

Vehicle systems

Can access online as well as stored information
Currently working to improve
The Future of H-C Dialogue
Systems

The final end result of human computer dialogue systems:

Seamless spoken interaction between a computer and a
human

This would be a major component of making an AI that can
pass the Turing Test

Be able to have a computer function as a teacher
Human Computer Dialogue in
Fiction


Halo's Cortana AI

Made from models of a real human brain

Made to run the ship

Made very human conversations
Ender's Game series: Jane

Made from "philotic connection"

Human conversation
Problems of Human-Computer
Dialogue

At the moment, most common computer dialogue systems
(call systems, chatter bots, etc.) cannot handle arbitrary
input

In many cases, the computer can only respond to "expected"
speech

Call systems often compensate with "Sorry, I didn't get that,"
when something unexpected is said.
Problems of Human-Computer
Dialogue

Computers need to be able to learn and process colloquial
speech

Needed to understand informal speakers:


Understanding varied responses for call systems

Accounting for variations in spoken numbers
Processing colloquialisms is also necessary for seamless
dialogue, where the computer must avoid sounding too
formal

John Connor: "No, no, no, no. You gotta listen to the way
people talk. You don't say 'affirmative,' or [stuff] like that. You
say 'no problemo.' "
Successes of HumanComputer Dialogue



So far, human-computer dialogue has been most successful
in applications where information about a specific topic is
sought from the computer.

Electronic calling systems: company-specific

Travel agents: specific to an airline or destination
However, more complex systems of human-computer
dialogue have been produced which can interpret more
varied input.

Physics tutoring system (ITSPOKE) which can analyze and
explain errors in the response to a physics problem.

Allows for more complex input than "Yes," "No," or "Flight UA93"
These still cannot compare to true human-human
dialogue.
Machine Translation

Important for:

accessing information in a foreign language

communication with speakers of other languages

The majority of documents on the world wide web are in
languages other than English

Google Translate

Bing Translate

WorldLingo
Statistical Translation

Rule based

Works relatively well with large sets of data

Used probability to translate text

Natural translations

Google
Example Based Translation

Converts "parallel" lines of text between language

Only accurate for simple lines

Analogy based
Future of Machine Translation

Goal:

Aim to be able to flawlessly translate languages

Link Human-Computer Dialogue and Machine Translation

Have someone be able to talk in one language to a
computer, translate for another person

Translated Video Chat
Machine Translation in Fiction


Star Wars: C-3PO

Interpreter

Could hear and translate alien languages

Final goal of machine translation
Star Trek: Universal Translator

Computer can seamlessly translate alien languages
Problems

Works well only with predictable texts.

Doesn't work well with domains where people want
translation the most:

spontaneous conversations

in person

on the telephone

and on the Internet
Problems


Computers can't deal with ambiguity, syntactic irregularity,
multiple word meanings and the influence of context.

Time flies like an arrow.

Fruit flies like a banana.
Accurate translation requires an understanding of the text,
situation, and a lot of facts about the world in general.
Problems

The sign is describing a
restaurant (the Chinese
text, 餐厅, means
"dining hall").

In the process of
making the sign, the
producers tried to
translate Chinese text
into English with a
machine translation
system, but the
software didn't work,
producing the error
message,
"Translation Server Error."

The software's user didn't
know English and thought the
error message was the
translation.
Successes

Product knowledge bases need to be translated into
multiple languages

Hiring a large multilingual support staff is expensive

Machine translation is cheaper and accurate with
predictable texts.

Microsoft, Apple, Google, Autodesk, Symantec, and Intel
use it.

Makes customers happy

Still readable though slightly chunkier than human translations
Assignment # 6
 Give
Presentation on any one of
the following projects
 Apple
Sri
 Google
Now
 Microsoft
Cortana
Questions???