Natural Language Translation (NLT) Task

A Natural Language Translation (NLT) Task is a language translation task that converts a linguistic expression into a semantically equivalent linguistic expression in some other natural language.

• Context:
  • Input: a Linguistic Artifact, a Natural Language.
  • output: an NL Translation Output.
  • It can range from being Manual NL Translation to being Semi-Automated NL Translation to being Automated NL Translation (solved by a language translation system.
  • It can range from being:
    • Text Translation, ...
    • Speech Translation, ...
    • Image Translation, ...
    • Video Translation, ...
  • It can range from being a Offline NLT Task to being a Real-Time NLT Task.
  • It can range from being a Forward NLT Task to being a Back-Translation NLT Task.
  • It can be supported by a Language Understanding Task, a Technical Term Translation Task, ...
  • ...
• Example(s):
  • Text Translation, such as Text Translate(“this sentence is written in English.”, Spanish).
  • Speech Translation, such as Video Translate(speech.mp3, Spanish).
  • Video Translation, such as Video Translate(video.mp4, Spanish).
  • …
• Counter-Example(s):
  • a Computer Program Translation Task.
  • a Paraphrasing Task.
  • a Language Generation Task, such as a summarization task.
• See: Lexical Semantics Task, Handwriting Recognition.

---

---

References

2009

• http://en.wikipedia.org/wiki/Translation
  • Translation is the interpreting of the meaning of a text and the subsequent production of an equivalent text, likewise called a "translation," that communicates the same message in another language. The text to be translated is called the source text, and the language that it is to be translated into is called the target language; the final product is sometimes called the target text.
  • Translation must take into account constraints that include context, the rules of grammar of the two languages, their writing conventions, and their idioms. A common misconception is that there exists a simple word-for-word correspondence between any two languages, and that translation is a straightforward mechanical process; such a word-for-word translation, however, cannot take into account context, grammar, conventions, and idioms.
  • Translation, when practiced by relatively bilingual individuals but especially when by persons with limited proficiency in one or both languages, involves a risk of spilling over of idioms and usages from the source language into the target language. On the other hand, inter-linguistic spillages have also served the useful purpose of importing calques and loanwords from a source language into a target language that had previously lacked a concept or a convenient expression for the concept. Translators and interpreters, professional as well as amateur, have thus played an important role in the evolution of languages and cultures.
  • The art of translation is as old as written literature. Parts of the Sumerian Epic of Gilgamesh, among the oldest known literary works, have been found in translations into several Asiatic languages of the second millennium BCE. The Epic of Gilgamesh may have been read, in their own languages, by early authors of the Bible and of the Iliad.
  • With the advent of computers, attempts have been made to computerize or otherwise automate the translation of natural-language texts (machine translation) or to use computers as an aid to translation (computer-assisted translation).

2008

• (Koehn, 2008) ⇒ Philipp Koehn. (2008). “Statistical Machine Translation." Cambridge University Press. ISBN:0521874157
  • The task of translation is to take words and sentences in one language and produce words and sentences with the same meaning in another language. Unfortunately, meaning is a very difficult concept to deal with.

---