Machine Translation

1
Machine Translation

• History of Machine Translation
• Difficulties in Machine Translation
• Structure of Machine Translation System
• Research methods for Machine Translation

2
History of Machine Translation

• Early 1950s
• Infancy time for Machine Translation
• Research was modest constrained by the
  limitations of hardware, esp. memories and slow
  access to storage
• A large number of projects supported by US
  agencies

3
History of MT (cont.)

• At the end of 1950s
• A bad time for Machine Translation
• Some promise before had remained unrealized
• The number of lab working in the field was
  sharply reduced all over the world
• Few of them were able to obtain funding for more
  long-range research programs

4
History of MT (cont.)

• A resurgence in the 1980s
• Many of the efforts were rapidly deemed
  successfully
• The most conspicuous example was the METEO
  system, developed at U. of Montreal
• METEO provided the French translations of weather
  reports used by airline, shipping companies and
  others

5
Difficulties in MT

• Some typical factors contribute to the difficulty
• Words with multiple meanings
• Sentences with multiple grammatical structures
• Uncertainty about what a pronoun refers
• Other problems of grammar

6
Difficulties in MT(cont.)

• Two common misunderstandings make translation
  seem simpler than it is
• Translation is not primarily a linguistic
  operation
• Translation is not an operation that preserves
  meaning
• Examples later

7
Difficulties in MT(cont.)

• Example for the first misunderstanding
• The police refused the students a permit because
  they feared violence.
• It is to be translated in to French
• police is feminine, they will also have to be
  feminine.
• Replace the word fearedwith advocated
• they refer to students and is masculine now

8
Difficulties in MT(cont.)

• For the second misunderstanding
• Different languages have different usage, for
  example there are languages like French in which
  pronouns must show number and gender, Japanese
  where pronouns are often omitted altogether,
  Russian where there are no articles, Chinese
  where nouns do not differentiate singular and
  plural nor verbs present and past

9
Structure of MT Systems

• Generally they all have lexical, morphological,
  syntactic and semantic components, one for each
  of the two languages, for treating basic words,
  complex words, sentences and meanings

10
Structure of MT Systems(cont.)

• transfer component the only one that is
  specialized for a particular pair of languages,
  which converts the most abstract source
  representation that can be achieved into a
  corresponding abstract target representation

11
Structure of MT Systems(cont.)

• Some systems make use of a so-called
  interlingua or intermediate language
• The transfer stage is divided into two steps, one
  translating a source sentence into the
  interlingua and the other translating the result
  of this into an abstract representation in the
  target language

12
Research methods of MT

• 1950s and 1960s
• Information theory, categorial grammar,
  transformational-generative grammar, dependency
  grammar
• 1970s and 1980s
• Artificial intelligence, non-linguistic knowledge
  bases, semantics and interlingua-based system

13
Research methods of MT(cont.)

• 1990s
• Neural networks, connectionism, parallel
  processing and statistical methods
• In the future
• MT research will be oriented towards the
  development of translation modules to be
  integrated in general office systems, rather
  than the design of systems to be self-contained
  and independent