The Difference Between Machine Translation Vs Computer assisted Translation (CAT).

The Difference Between Machine Translation Vs Computer assisted Translation (CAT)

 

There is a significant difference between the two, with very different results. The terms “computer-assisted translation” and “machine translation” sound similar, and it’s easy to get them confused.

Machine Translation: Fast and Cheap, but Inaccurate

   Machine translation is accomplished by feeding a text to a computer algorithm that translates it automatically into another language. That is, no human is involved in the translation process.

   The advantages of machine translation include cost and speed. Computers can process a machine translation almost instantly. There are free programs such as Google Translate that can translate relatively short texts instantly, but if you need to translate a very long document, you can purchase software that can process an unlimited amount of text at the cost of the software alone. There is also software available that can be integrated with other computer and online tools, providing instant translations in various contexts.

    The major disadvantage is lack of accuracy. If you’ve ever used Google Translate to attempt to understand a text in a foreign language, you will know that this method does not produce a particularly natural-sounding or accurate translation. Language is highly complex and dynamic, and while this type of translation technology has improved greatly over the years, it will never be able to completely accurately identify the nuances of each language and transfer them into another language.

   It is possible to hire a “post-translation editor” to look over the translation and correct errors, but it can be harder to correctly deduct the meaning of a sentence from its machine translation than from its original language. Translators hired to “smooth out” such translations sometimes end up asking clients to send them the original text because the translation was unintelligible. This is a big waste of everybody’s time!

   The best use for machine translation, then, is when you need to understand the general gist of a text. If you need an accurate translation that anyone can understand, you’ll want to opt for a computer-assisted translation.

 

Computer-Assisted Translation(CAT): Human Translation Enhanced with Computerized Tools

   Computer-Assisted Translation is a human translation carried out with the aid of computerized tools. That is, a human translator is the one reading and deducing the meaning of the source text and transferring it into the target language. They are simply utilizing computerized translation tools to help them work more quickly and accurately.

   You probably already use some of these tools yourself. For example, nearly every word processor, and many web browsers, have a built-in spell checker and/or automatic spelling correction function. This saves writers and translators a lot of time looking up words in the dictionary!

   Speaking of dictionaries, when a translator does need to look up a word, they can save time by using a computerized dictionary. As a translator, my most often-used tools are the multi-language dictionary (to help recall words that may be escaping me at that moment) and the thesaurus (to help me choose exactly the right word for my translation).

   More complex computerized translation tools include translation memory tools (databases of texts in multiple languages), terminology managers (that help translators maintain consistent terminology throughout the translation), terminology databases (to help translators locate the correct terminology for that field), bitext aligners (which align the source text and the translation for side-by-side comparison), and more.

 

Types of Computer-Assisted Translation(CAT):

1. TRANSLATION MEMORY SOFTWARE

    Translation memory software is the most well-known CAT tool. It divides the texts to be translated into units called “segments”. As the translator advances in the translation of the document, the software stores the text in a database of already translated segments. When the software recognizes that a new segment is similar to a segment already translated, it suggests that the translator reuse it. Some translation memory programs do not work with databases created during a translation, but with preloaded reference documents.

Some examples of translation memory software: Trados Workbench, DéjàVuX, SDLX, Star Transit, MultiTrans, Similis, MetaTexis.

 

2. LANGUAGE SEARCH-ENGINE SOFTWARE

    Linguistic search engines work like traditional search engines, except that they do not seek results on the Internet, but in a large database of translation memory. The goal is to find, in these banks, fragments of previously translated texts that match the new text to be translated. Linguee, a multilingual context dictionary, is one of them.

 

3. TERMINOLOGY MANAGEMENT SOFTWARE

     Among CAT tools, there is also terminology management software. With programs of this type, the translator has the ability to automatically search for the terms in a new document in a database. Some of these systems allow the translator to add, in the database, new pairs of words that match and verify text using various functions: the translator can then check whether this or that term has been translated correctly and consistently throughout the whole draft. Here are three examples of this type of software: SDL MultiTerm, LogiTerm and Termex.

 

4. ALIGNMENT SOFTWARE

    Text alignment programs allow the translator to build a translation memory using the source and destination of the same text: the software divides the two texts into segments and attempts to determine which segments agree with each other. The result of this operation can be imported into a translation memory software for future translations. Here are four examples of alignment software: Bitext2, Tmx Bligner, YouAlign and LF Aligner.

 

5. INTERACTIVE MACHINE TRANSLATION

    Automatic interactive translation resembles the programs you use on your cell phone for writing messages: the program tries to predict how the human translator would translate a phrase or sentence fragment.

 

OTHER LANGUAGE PROGRAMS OF HELP TO THE TRANSLATOR

Finally, you should also consider other very useful linguistic software for translators:

Spell checkers (Proofread).

Grammar checkers (Grammarly, Reverso)

Terminology databases or online dictionaries, such as TERMIUM Plus, and the IATE.

Search tools for “full text” and indexing which allow searches to be carried out into already-translated texts or reference documents of all kinds, such as for example, ISYS Search Software and dtSearch Desktop.

Concordant or matching software which are reference tools used to look up a word together with its context, whether in a monolingual, bilingual or multilingual body (such as a bitext or a translation memory).

Project management software. With this program, a project manager at a translation company can organize complex projects by assigning translation tasks to different translators and track the progress of each one.

 

Types of Machine Translation:

1.     Rule-Based Machine Translation (RBMT)

RBMT, developed several decades ago, was the first practical approach to machine translation. It works by parsing a source sentence to identify words and analyze its structure, and then converting it into the target language based on a manually determined set of rules encoded by linguistic experts. The rules attempt to define correspondences between the structure of the source language and that of the target language.

The advantage of RBMT is that a good engine can translate a wide range of texts without the need for large bilingual corpora, as in statistical machine translation. However, the development of an RBMT system is time-consuming and labor-intensive and may take several years for one language pair. Additionally, human-encoded rules are unable to cover all possible linguistic phenomena and conflicts between existing rules may lead to poor translation quality when facing real-life texts. For example, RBMT engines don’t deal well with slang or metaphorical texts. For this reason, rule-based translation has largely been replaced by statistical machine translation or hybrid systems, though it remains useful for less common language pairs where there are not enough corpora to train an SMT engine.

 

2.    Statistical Machine Translation (SMT)

SMT works by training the translation engine with a very large volume of bilingual (source texts and their translations) and monolingual corpora. The system looks for statistical correlations between source texts and translations, both for entire segments and for shorter phrases within each segment, building a so-called translation model. It then generates confidence scores for how likely it is that a given source text will map to a translation. The translation engine itself has no notion of rules or grammar. SMT is the core of systems used by Google Translate and Bing Translator, and is the most common form of MT in use today.

The key advantage of statistical machine translation is that it eliminates the need to handcraft a translation engine for each language pair and create linguistic rule sets, as is the case with RBMT. With a large enough collection of texts, you can train a generic translation engine for any language pair and even for a particular industry or domain of expertise. With large and suitable training corpora, SMT usually translates well enough for comprehension. The main disadvantage of statistical machine translation is that it requires very large and well-organized bilingual corpora for each language pair. SMT engines fail when presented with texts that are not similar to material in the training corpora. For example, a translation engine that was trained using technical texts will have a difficult time translating texts written in casual style. Therefore, it is important to train the engine with texts that are similar to the material that will be translated.