Machine Translation

Outline
 Introduction
 Types of MT
 History of Machine Translation
 Existing Machine Translation System
 Approaches to MT
 Translation Unit

March 12, 2025 2

Introduction to Machine
Translation
Translation?
 The process of transforming natural language
text from one language (source language) into
a meaningfully equivalent text in the second
language (target language) is called
translation.
 Ref: [M.A Khan (1995)]
Source Target
language language
Translation by human
being
 Source text ----------------> Target text
March 12, 2025 4
Translation…
 Note:
 The closeness of the target text into the
source text is an idealism which is not
always achievable.

 The closer the meaning of the two text

to each other, the more accurate will be
the translation and the higher will be the
quality of translation.

March 12, 2025 5

Human translator
 A person who is responsible for the translation
of one natural language text (or speech) to
another natural language text (or speech) is
called human translator

 Also called technical translator

March 12, 2025 6

Types of human translator
 Language specialist:
A specialist of some field, e.g. engineering etc.,
but having specialties in one or several foreign
languages
 Linguists:
 Know more than one languages from the very
beginning
 A specialist of languages
 Trained person:
 Geteducation of more than one languages in
some educational institute
March 12, 2025 7
Problems with a human translator
 Shortage
 Can get tried and bored
 Expensive/Charge fee
 Not available everywhere
 Can die
 Translate same word/sentence differently in
different times i.e. inconsistent.

March 12, 2025 8

Need of a machine translation system
 To overcome the limitations of human
translators, we need machine translation
systems

March 12, 2025 9

Machine translation…
 Definition 1:
 Machine Translation (MT) is the task of
automatically converting one natural language
into another, preserving the meaning of the
input text, and producing fluent text in the
output language.
 Ref: [http://nlp.stanford.edu/projects/mt.shtml,]

Source Target
language language
Translation by
 machine
Source text ----------------> Target text
March 12, 2025 10
Types of MT
Types of MT

 Direct/ Dictionary based MT

 Transfer-based MT
 Interlingual MT

March 12, 2025 12

Direct or Dictionary based MT
 The words are translated as a dictionary does i.e.
word by word usually without much correlation of
meaning between them.
 Dictionary lookups may be done with or
without morphological analysis
 While this approach to MT is probably the least
sophisticated, dictionary-based machine
translation is ideally suitable for the translation of
long lists of phrases e.g. inventories or simple
catalogs of products and services.
March 12, 2025 13
Transfer-based MT
 It is currently one of the most widely used methods of MT.
 Transfer MT breaks translation into three steps:
 analysis of the source language text to
determine its grammatical structure,
 transfer of the resulting structure to a structure
suitable for generating text in the target
language,
 finally generation of the text.
 Transfer-based MT systems are thus capable of using
knowledge of the source and target languages

March 12, 2025 14

March 12, 2025 15
Interlingual MT

 One of the classic approaches to machine

translation.
 The source language is transferred into an
interlingua, i.e., an abstract language-
independent representation.
 The target language is then generated from the
interlingua.

March 12, 2025 16

History of MT
History of machine translation….
 The early years 1950s—1960s
 Warren Weaver proposal
 The first proposals for machine translation using computers were put
forward by Warren Weaver, in his July 1949.
 Georgetown experiment:
 The first work of translation was published in 1954 in the Georgetown
experiment involving fully automatic translation of more than 60
Russian sentences into English.
 The experiment was a great success and the authors claimed that
machine translation would be used in translations within three or five
years.
 However, the real progress was very slow

 Ref:[http://www.thelanguagetranslation.com/machine-translation.html
]

March 12, 2025 18

History of machine translation….
 1966-- Cut off funding
 Reasons for giving off research in MT
 Expensive memory
 Low size of memory
 Slower processing speed of processor
 Low quality softwares to process natural language text
 Large amount of time needed to processes limited text
 Result
 Funding was stopped for further research on machine
translation projects

March 12, 2025 19

History of machine translation....
 1970 and early 1980: Revival of machine translation
 Development of personal computer:
 However, starting in the late 1970s and beginning 1980s, with the impact of
personal computer revolution, with the increase in computational power, more
interest began to be shown in statistical models for machine translation.
 Large and less expensive memory and computers:
 There was growth in the use of machine translation as a result of the beginning
of less expensive , large memory size and more powerful computers.
 In 1970, the Systran system was installed for the United States
Air Force and subsequently in 1976 by the Commission of the
European Communities
 The METEO System, was installed in Canada in 1977 to
translate weather forecasts from English to French

 Ref: [http://www.thelanguagetranslation.com/machine-translation.html]

March 12, 2025 20

History of machine translation....
 1980s
 Research during the 1980s typically relied on
translation through some variety of
intermediary linguistic representation involving
morphological, syntactic, and semantic
analysis.
 Statistical Machine Translation (SMT) approach
was introduced by IBM
 Example-based Machine Translation (EBMT)
approach was introduces by Makoto Nagao.

March 12, 2025 21

History of machine translation....
 Early 1990s
 Speech translation
 During the 1990s, research began into speech
translation after the successes in speech recognition
and speech synthesis
 Online translation systems
With the 1990s, the importance of machine translation
further increased and the use of "translation engines" on
the Internet to allow for translation of websites and email
languages started
 Ref: [
http://www.thelanguagetranslation.com/machine-translation.html]

March 12, 2025 22

History of machine translation....
 Examples of online translation systems
 SYSTRAN
 http://www.systranet.com/translate
 Google translate
 http://www.translate.google.com
 AltaVista's Babelfish
 http://babelfish.yahoo.com/
 Microsoft translator
 http://www.microsofttranslator.com/

March 12, 2025 23

History of machine translation….
 Recent research: beyond 1990s
 Work is continued in:
 statistical machine translation
 example-based machine translation
 speech translation
 research has focused on moving from specific-domain systems to
generic-domain translation systems
 In the United States (STR-DUST) solutions for automatically
translating Parliamentary speeches and broadcast news have been
developed
 More recently, the French-German project Quaero investigates
possibilities to make use of machine translations for a multi-lingual
internet. The project seeks to translate not only webpages, but also
videos and audio files found on the internet
 Ref [http://en.wikipedia.org/wiki/History_of_machine_translation]
March 12, 2025 24
Importance of machine translation
 Provides Cheap, universal access to world’s online
information regardless of original language.
 Used in education, politics, business, trade,
cultural study and media.
 A combination of text+speech translation
enables:
 Two persons, that do not know each other language, to
talk to each other on the telephone
 Participants in a conferences, debates and discussion
forum in international organization such as UNO, ECO
etc to talk to each other.
 The translation of films in film industry to other
languages
 To March
translate
12, 2025 a spoken or written language to the
25
Problems in machine translation
 Bilingual or multilingual dictionary design
 Poly-semantic words of the source language
 Problem of idiosyncrasies: no relevant word in target language
 Order of words of the source and target language
 No one-to-one correspondence between the words of the
source and target language

March 12, 2025 26

Existing Machine
Translation System
GETA
 Group study for automatic translation”.
 Initial name was CITA
 It was developed in 1959 at CNRS (Centre
national de la recherche scientifique) France
 Initial objective of this was to translate the
space messages of Russian into French

March 12, 2025 28

GETA (Cont..)

The translation process in GETA consist of three

main stages
 Analysis
 Transfer
 Generation

March 12, 2025 29

EUROTRA
 Initiated in 1978 by European Community
R&D program in MT
 The EUROTRA project is the machine
translation program of the European
Community
 The objective of this project is the creation of
a small multilingual system

March 12, 2025 30

EUROTRA (Cont...)
 Operated in a limited subject field and for a
limited no. of text types
 Served as a basis for industrial development
 Consist of three main phases i.e. analysis,
transfer and generation

March 12, 2025 31

DLT
 Distributed language translation
 Is a Dutch MT system
 It was initiated by a Dutch engineer Toon
Witkam in 1979
 Consist of two main phases i.e.
 Analysisand
 Generation

March 12, 2025 32

DLT (Cont...)

 This system use interlingua approach

 Translate text from English into French

March 12, 2025 33

SUSY
 This MT system was developed in the
University of Saar (West Germany)
 Based on the principles of multilinguality
 Initially work began on Russian and German
languages

March 12, 2025 34

SUSY (Cont...)
 The system consist of three major components
 Analysis
 Transferand
 Generation
 Nowadays, SUSY is called STS
 Use for computer-aided translation

March 12, 2025 35

SYSTRAN
 Developed in 1960s by Peter Toma
 English into French and Italian
 It was first used by U.S. Air Force
 1988 it was working for 12 languages
 This system is “hybrid ‘direct transfer’
system”
 Means the system uses partly transfer strategy
and partly direct
March 12, 2025 36
BABEL-Research
 It is a group of researchers in the University of
Brussels, Belgium
 Funded by UNISYS
 Translation of technical manual from one language
into another
 The languages involve are English, French, Dutch
and German
 Transfer architecture is used
 The programming language use is Pascal

March 12, 2025 37

SYSTRAN system

March 12, 2025 http://www.systranet.com/translate 38

Google translate system

https://translate.google.com.pk/

March 12, 2025 39

Example

March 12, 2025 40

Yahoo Babel Fish system

March 12, 2025 41

Microsoft translator

March 12, 2025 42

Other online machine translation systems

March 12, 2025 43

March 12, 2025 44