google translate

anyone have an idea on when google translate will be perfected? I ask because if and when that happens, It will be my dream translator.

 

Log in or Sign up to hide all adverts.

I use google-translation and many times is stupid.
Some words are translated inaccurate
and also the location of each word in the sentence is incorrect.

Please Register or Log in to view the hidden image!

 

Log in or Sign up to hide all adverts.

I'm not sure, but I can say Google translator seems to have very long distance to go. I'm a Korean, and I sometimes use Google translator. I think its performance has been being improved gradually. However, I find that the translator is very poor at making sentences in a sense that native Korean speakers understand, whenever I use it.

 

Log in or Sign up to hide all adverts.

Meanwhile, it's the best translator out there, but it's still too much inhuman.

 

Translation programs tend to have issues with Syntax, this is greatly increased by a persons own individual method of expression. As one person might right one sentence a particular way, and another will write it completely differently.

An example:

• I made haste while walking to the shop.
• To the shop I made haste while walking.
• I walked to the shop as fast as I could.

To an English speaker these sentences all attempt to suggest the same thing, however after translation they can become extremely different, which is one of the main problems with dealing with a second or more language. Sentences can be read in many different ways, translators attempt to use pre-defined syntax structures however they will be defined by one person or a group of key persons, as opposed to the nature of language itself being a twist of individuality and localised accenting.

 

Progress in machine translation amazing, has been. Voice recognition too. Practical universal translators in our portable devices are coming soon, and this technology is going to change things better much. Lose hangups of syntax we will, learning thinking how others.

 

Roughly 4-5 years.

Right now the biggest breakthroughs in language processing involve using AI to statistically infer the meaning of words. The hardware required to do this robustly and quickly is still fairly cutting edge. See: Watson competing on Jeopardy.

For example, suppose you say "I'm glad Los Angeles won." The average person will do a word association (i.e, search for a statistical correlation) and interpret a sentence probabilistically:

Los Angeles + win -> competition -> sports team -> Lakers winning 2010 finals -> You are glad the Lakers won the 2010 championship.

For a computer to do this is fairly tough. Search engines can mimic this, but they're not really interpreting syntax as much as they are optimizing search results based on a ton of feedback. You are beginning to see basic syntax interpretation with add-ons like True Knowledge.

A big sign we're getting close to real-time translators is when a computer will successfully punctuate a variety of sentences. That's a sign they're "getting it." Once it can do that, reliable speech recognition and translation is will be imminent.

The real importance of this won't be translation between two languages like English and Korean. It will be the translation between everyday natural language and computer language: that is, you can tell your computer to do a task in natural language and it will respond. Today people program instructions for computers. In ten years, they will delegate them.

 

.

try yamli translator, it's alot better fir the main languages.

www.yamli.com

 

nope even that still has a ways to go. I tested it out by doing this, having it translate I simple Spanish sentence two different ways. Yo quiero amar and quiero amar. To a Spaniard these to sentences mean the same thing. "I want to love" but yamli didn't have the same answer for both. Btw same goes for google translate.

 

I hope it only takes that long.

this has already been done by a program called Dragon Naturally Speaking.

 

Actually, I don't think programming will be done by Dictation alone. Dictation is still very linear and long-winded, It's more likely that apparatus will be used to read the human mind in regards to it's various wavelengths and overall brain patterns, and those outputs will then be converted through macro's into programmatic actions.

It will be much like virtual reality modelling, however with the capacity for the human brain to interact at a "Kernel Debugging level" with any programming project that's undertaken.

 

.

ah, well, internet translators are not all so good, try trasnlating from english for example, to japanese, itwill be a disaster!

Please Register or Log in to view the hidden image!

but from japanese to english is fine

 

Google translate sucks, its Japanese and Chinese translations suck ( I can read kanji almost perfectly, so I know ).

Here is an example:

吾皇萬歲，萬歲，萬萬歲

Check it for both Japanese and Chinese on translator ( they mean pretty much the same thing in both languages ), afterward I will tell you the right translation. ( Wikipedia got it wrong as well ).

 

Dude, Japanese to English still sucks, but not as much as vice versa.......