Slashdot Mirror
Distributed Translation Project
moon unit beta writes "New Scientist has this story about a new plan to build a multi-language translation database called the World Wide Lexicon, using a distributed community of volunteers. The designer compares it to a distributed computing project and believes it could make it easier to translate more obscure languages."
The Universal Translator is finally here! But will they ever release it in fish form?
Everyone translate the word "fuck" into your native language.
If tits were wings it'd be flying around.
So, I can use a plugin that would automatically use this super dooper distributed brain to get all my french pages into english etc?
:D http://www.pornolize.com/
Currently my favorate web translator is this one
----- Whats wrong with this picture? http://www.revoh.org:1234/whatswrong
"This will automatically detect when the computer user is less busy and ask them to translate a word or phrase."
i wonder how its gonna detect when the user is not busy. this software can never be installed on something like my home computer where i leave my DSL on to make it work on SETI.
What's in it for the volunteers? Seems that novelty might bring experts in to volunteer short term, but when businesses, academics, etc. begin using the service in volume, it really will cry out for commercialization. The volunteers won't stick around performing translations gratis forever. At some point you have to pay them per translation or provide some other compensation (perhaps a /. like karma system?)
The related bigger question will be whether this model ultimately proves to deliver quality translations at a lower cost than a traditional translation service. I don't see how this could happen if you have to still have a language expert look at the full translation as a whole to ensure that contextual subtleties are not lost.
Great! Now we'll have Engrish resulting not just terrible Japanese->English translation, but all kinds of other languages too. Eventually the web will be so filled with bad grammar that the next generation will have no idea how to string a simple sentence together. Looks like we will have to start compiling our correspondance after all... for coherence.
Liora
[snip]"One of the main problems is quality assurance," says Ramesh Krishnamurthy, a linguistics expert at the University of Wolverhampton, in the UK. "Translation is a highly developed skill." [snip] But Paul Rayson, a research fellow at Lancaster University, adds that unskilled translators may confuse the meaning of individual words. "The problem is you generally need the context to get a good translation," he says.[snip]
This looks like it will be a very cool project, but for corporate/buisiness use I don't think it would ever fly.
If you have ever played in the area of i18n then you will quickly understand why this pbly won't work perfectly. There are so many caveats to each language, tone, context etc... This might be a useful starting point for transaltion services, but for the final cut, it would still need to be checked and double checked by a translation service.
I still think its very cool though ;)
-ryanI'd also like to applaud them finally including the lost language of Ur in their translation project. For too long the ancient Sumerians have been excluded from contributing to the global society due to their lack of knowledge of English, French, Spanish, Swahili or Chinese.
Where can I download the screensaver so that I can contribute?
More people speak Klingon than Navaho...
It's a matter of days until someone will request a log of people connecting to the server during work-hours... Here is the beauty of the seti@home client: computers can have spare cycles, people don't.
-- No sig today
If it's going to detect when I'm "less busy." Is this going to pop up a window in my face every time I spend more than a couple minutes mentally composing prose or code? The potential for user annoyance here seems incredibly high to me...
Distributed computing is an elegant and efficient use of otherwise untapped resources--cycles that are literally "going to waste" (in one sense). By hitting up the users, though, you're attempting to use a resource that is anything but untapped: that user's time. It might work, but let's not bill this as anything other than what it is--asking for volunteer work from people.
Which isn't really that new an idea.
Reality has a conservative bias: it conserves mass, energy, momentum...
One of the big issues with translating between human languages is context. While many words have more or less direct equivilants in other languages ("dog"(en) "perro"(es)), you're always going to run into slang, cultural references, and especially, jargon, where the particular usage will not be in a standard dictionary, and only by the context can the actual meaning be inferred (Example: the word "anchor" in the context of sailing versus the context of webpage design).
Not that this can't be overcome with the distributed model the article discusses, but I still think it will be a while before we see computer translation that doesn't require at least some degree of human assistance.
Is there some way to translate into a common universal "intermediary" language, then translate to the destination language?
I'm just thinking that most languages could relate more closely with an "iconographic" type language than with the idiosyncrosies of other languages. For concrete ideas this may work well, but for more conceptual ideas this may fall apart...
Just my $0.02, being uneducated in linguistics...
MadCow.
I used to have a sig, but I set it free and it never came back.
I send you this words in order to have your translation
I'm not a translator but during college I worked with a comparative lit professor who translated novels from spanish into english. The problem with translation is wrestling with the subtle shades of meaning that every single word has and to find its perfect pair in the language you're translating into. Then you have to adress the context in which the word was written (the larger sentence--what information is it trying to convey, what mood (much trickier) is it trying to imply, and finally does this match the author's style and the novel's tone (this is what truly makes translation an art).
This is a bad example but just so you get the idea, it's hard even english to english:
original:
John hurried to the shopping mall.
variants:
John made great haste to get to the shopping centre.
John ran to his destination, the shopping mall.
John rushed to the store.
John spared not the whip in perambulating to the suburban commericial district.
John ran off to waste time at the corporate copyright paradise.
blah blah blah...
From the article:
"The problem is you generally need the context to get a good translation,"
This is very, very true. Any competent translator can tell you that it's almost impossible to get a fully accurate translation from just a few lines or words... context is absolutely imperative. This looks a lot like vaporware to me.
And then what about when the smart-ass teenaged year old kid signs up, gets bored and starts translating to obscene or nonsensical results? They'll need some sort of moderation system, if this is to work at all.
Thanks, newscientist, for bringing us another well researched and peer-reviewed story, maintaining the image that a "new scientist" is one who has forgotten about the scientific method.
--Use this space for notes--
Hello,
I am the lead developer working on the WWL project. There are actually two components to this project. Overall, the NS article did a good job of explaining it, but it was based on a phone interview so some material got lost in translation, no pun intended.
There are two components to the project.
1. One is a simple SOAP based protocol (WWLP) that will be published soon, in early May. This protocol creates a standard set of methods for discovering and communicating with existing dictionary and semantic network servers (of which there are many).
Think of this as GNUtella for dictionaries. A WWLP aware program starts up, invokes a SOAP method to a supernode to locate Russian-Spanish dictionaries. Then, it contacts one or more of these dictionaries to search for words, synonyms, etc.
The basic goal is to standardize the client/server interface for dictionaries. They all provide the same basic services, but have slightly different front ends. So just doing this will make it easy to incorporate dictionary functions into many types of apps (and also make existing dictionaries more visible to internet users).
The idea is similar to an older TCP based protocol called DICT, except that it is easy to implement in high level languages, SOAP aware scripting languages, etc. It also provides a discovery mechanism so you can automate the process of finding an Urdu-English dictionary for example.
2. The distributed computing (or distributed human computing) project. The NS article mainly focused on this. The idea here is to enlist a large number of internet users to help build and maintain a dictionary (which will also be visible through the WWLP interface).
The goal here is to create a mechanism for collecting definitions and translations for words and phrases in less common language pairs (as well as for slang terms that are not covered by most formal dictionaries).
....
The goal in both cases is to make it easy to find and use dictionary services throughout the web, and create an incentive for people to build their own dictionaries. This is NOT a translation system, although it can be incorporated into translation software (for example, to extend the number of words covered).
Thanks for your time.
Brian McConnell
PS - if you want more information, check out www.worldwidelexicon.org
Yes, you can do a word-for-word translation of most words in any language. No, you'll need a very sophisticated system to get the meaning to a reader.
The main problem is that sentence structures are different, idioms get in the way, and words have more than one meaning. A human translator has the power to take a set of words, convert it to an idea, and put out a different set of words, something no machine can do.
Here's a lamebrained example: "The spirit is willing but the flesh is weak." Convert that to Russian and back and you might get, "The liquor will do it but the meat is bad." For a hands-on example, try converting the first few paragraphs of a news article into French using The Fish. On a personal note, I had a conversation with a German guy on ICQ once, using the fish. The results were...interesting. I also read Indonesian newspapers, and I assure you that a literal translator would hurt itself quite badly on this...let alone a less English-like language such as Arabic or Japanese.
That being said, why not use distributed human computing for the thing it's good at? Instead of translating words, how about sentences? You can get at the ideas much better this way. Those sentences that hadn't been translated yet could show up as literal words; those words that hadn't been translated would show up natively. I mean, if you've got human translators for this, you can do things that are not restricted to computers. I can think of a lot neater things the guy proposing this can do with this idea than what he's come up with so far.
There's no sig like this sig anywhere near this sig, so this must be the sig.
You need a lot of "mod" and "metamod"-like activities to work; it looks to me that the peer review system shouldn't be too "democratic" to succeed (i.e., there is always a need for some top-level superusers, who are trusted automatically because they are essentially the system builders).
Anyone has an example of such a system with its founders going berserk (say, think of CmdrTaco starting daily trolling :-) )?
VKh
However, this will do nothing to aid in machine translation. You can't simply translate individual words from one language to another, or even short phrases. Translators such as Babelfish understand the basic rules of grammar in each language in order to handle fundamental differences in the way different languages put sentences together.
But Babelfish and other online translators are still a far cry from doing true translation, because they don't understand the text they're trying to translate.
When a machine generates a translation, there are no issues of copyright ownership, because machines are not authors in the statutory sense; the owner of the machine can claim copyright and move on.
When individual human translators get involved, there's an entirely different order of complication. Sure, it's possible to use licenses like the OPL (Open Publication License) to navigate these complications, but the compliance problems remain an obstacle to overcome. It'll be tough to remain competitive when babelfish and google don't have to put up with similar issues.
When this is added to all the other problems associated with massively distributed activities relying on humans to function, I just can't see how it'll succeed. Too bad, perhaps, but nonetheless true.
From the orignal source (http://picto.weblogger.com)
While the SETI At Home Project taps the idle CPUs of millions of personal computers, the worldwide lexicon enlists the help of internet users who are logged in, but not chatting. Think of this as distributed human computation.
"Distributed human computation"? Is that like using up all those spare brain cells you weren't using right now?
then you should go to their site, which was completely unmentioned in the article: wwl page
visit the hwky website for a lyrical genius infusion.
It doesn't work. If translating were so simple for machines to do they'd be doing a fine job. However good translation requires context, insight, emotional inflection, etc. Even then each and every one ends up different; sometimes subtly sometimes blatantly.
Just as machine translation sux at these so will distributed translation. Reading a paragraph or a page doesn't tell enough about the feel, flow, or tone of a document. There are numerous words and phrases that can be interpreted multiple ways between any two languages and will be, each time differently by each interpreter.
If you don't know this already then go and look up any document (books and short stories are easy to find, so is poetry) that has been translated more then once. Take a look at the different translations and ask yourself - "Are these really from the same source document?"
Now imagine trying to read something composed of alternating paragraphs or pages from each translation: Incoherence.
Distributed problem solving works for subjects with clearly defined data sets, methodologies, and standards; not human language.
I don't read ACs: If a post isn't worth so much as a nom de plume to its author then I wont bother either.
Way to go guys! All of the SlashTrolls know about it now too. What I thought I asked:
"Where is the restroom?"
What the native speaker heard me say?
"I want to slowly and lovingly take your wife in the rectum."
I recall a Monty Python sketch where a guy was put on trial for fraudulent phrasebooks that did that sort of thing. Someone gave the phrasebook guy a tainted phrasebook from his language back into english and he kept insulting the judge. Hilarious.
How far can we trust this translation project once the trolls make a few choice "contributions"?
I tried to post the translations themselves, but the "lameness filter" considered it too many "junk characters", even after I removed all the accents and umlauts and such. The lameness filter is lameness incarnate.
"I don't know half of you half as well as I should like, and I like less than half of you half as well as you deserve."
C'mon folks, this is a troll! Who the heck fell for it?!
I don't read ACs: If a post isn't worth so much as a nom de plume to its author then I wont bother either.
> Think about all the 12-year-olds -- script kiddies or not -- who will pretend to know a language and just type in a random collection of letters.
I dont know if you remember what it was like to be 12, but while I might have done what you'd proposed once, twice, I can't imagine the amount of 'noise' in this translation service coming from 12 years old who finally find their life long mischevious passion of offering 'bogus' translation services.
I mean, really, do you see 12 year olds downloading a distrbuted translation app, translating 'bogus'ly, and getting their jolies from this in any quantity that dimishes the value or effectiveness of this project? 12 year olds have much more important things to do, like learn how great masturbation is, and play videogames, and other forums where 'abuse' is fairly indistiguishable from proper use.
"Old man yells at systemd"
This sounds suspiciously like the idea that any unsolved computer science problem can be solved by adding another level of indirection.
If tits were wings it'd be flying around.
...a multi-language translation database called the World Wide Lexicon, using a distributed community of volunteers....
As soon as I read this, I immediately thought of Google's pigeon-based page-ranking technology. "I just hope those volunteers can type really really fast...."
ms
I used to believe in the whole idea of grammar. Until I went into Speech research and learned more about language. One of the big breakthroughs in speech recognition was when they went to hidden markov models for language. This language modelling technique is now used on all modern recogniziers is statistical in nature, not grammatical [rule based]. Grammatical models are never flexible or robust enough to represent true spoken speech.
The fact is that English is an organic language, and has organic properties. It grows. It changes. It has fuzzy boundaries. We must expect language constructions to change with time--it has been changing all along! All the rules and regulations you learned about grammar are generally context senstive, and do not hold up in all contexts, most notably, spoken speech. The rules of grammar are artificial, really imposed by publishers as a standard, but they do not actually reflect the full spectrum of the language.
You did.
Unless you are also a troll, in which case the answer would be me.
--
E_NOSIG
Trolling on /. most likely results from the very short amount of time it takes to see people responding to your crap. Most scipt-kiddie like behaviour is similar, when you start a DOS attack the results of your mischief is immediate. This translation service on the other hand will probably prove to be quite boring and thus only those with dedication will be able to commit to doing a translation instead of watching The Simpsons.
t.
The implications for quantum computing are overwhelming!
Bjarke Roune
Handling 'it' is quite easy.
t.
Lacklider also did a lot of papers on this in the 1960s, Xerox PARC did huge amounts of research and lots of other people including myself have worked (http://www.freesoftware.fsf.org/cdf/) on this problem. What I will say is that although lots of us have worked on it there are very few working systems (mind works a little but needs huge developement), if these guys succeed it is good for us all.
So good luck Wolverhampton!
e4 e5
t.
Technically, the multiple translations for snow/ice in Eskimo is a misnomer. Amerindian languages all have multiple words for almost everything.An example:
to cook has maybe 8 or 9 translations depending on the Amerindian Family...to cook fast, to cook raw meat, to cook fish, etc
My take on Amerindian languages being like this was for survival. There could be no doubt what the speaker was trying to convey. The snow thing with eskimo describes different kinds of snow, wet snow, slushy snow, snow that you could sled on, snow that a dog pissed on, etc. Tlingit (Washington State & Canada) has the same thing. The Tlingits are known for the Totem poles.They have different words for the verb "to fish"
If you're not a Liberal in your 20's, then you have no heart.If you're still a Liberal in your 30's you have no brain.
(Dang, left my flame suit at home. Oh, well.)
It seems like the creators of this system have noble goals, and I appreciate their efforts. It reminds me of Esperanto's Prague Manifesto. "Every language both liberates and imprisons its users, giving them the ability to communicate among themselves but barring them from communication with others."
I think anything that can bring the disparate world together is a good thing. But we woulnd't need technology like this if everyone got off their duff and learned a second language. For the purpose of learning a common second language, Esperanto is ideal. A smart kid like you can learn it in just a few hours of study.
I've used it to communicate with people from Brazil, Korea, and Germany, without having to learn Portuguese, Korean, and German. We just learned a simple middleware language to help us communicate. The Esperanto community offers Free Tutored Courses to help you get started. It's well worth the small investment to become bilingual.
But don't take my word for it. In the words of Tolkein: "My advice to all who have the time or inclination to concern themselves with the international language movement would be, 'Back Esperanto loyally.'"
-- Yekrats
Ceci n'est pas une pipe.
"I went to the store."
might become:etc.
Granted, the first markup pass would be a killer, but subsequent translations could be automated. As an added bonus, kids would get to learn grammar again.
(Definitions should really be a URI to a universal dictionary, but then you knew that...)
Serving your airship needs since 1995.
The easiest way would be if you have everything translate to one 'central' language, and from there have a reverse. That way you wouldn't need 1 two-way for every language each language had to contact (ie, 50 per language), but rather one two-way for each language. I think something would be lost, but this makes the project infinately easier to do, and to expand on (not having to write 50 programs to put in 1 more language).
In my opinion, the best approach (NOT best result), and the most likely to succeed.
There are supposedly 17 different meanings for "pretty little girls school", and that's before you get beyond syntax.
One interesting attempt at a language was Loglan. It had an computer grammar. It had regular syntax and simple phonetics. It was designed to be easy for anyone to learn (though significantly easier for English speakers, and secondarily fro other Indo-European language speakers). Unfortunately, to my mind what this clearly proved was that there was no good theory of semantics.
Well, that was 20 years ago. Perhaps the theories of semantics have improved since then. I haven't been watching. But I really do have my doubts.
I think we've pushed this "anyone can grow up to be president" thing too far.
The goal here is to create a mechanism for collecting definitions and translations for words and phrases in less common language pairs (as well as for slang terms that are not covered by most formal dictionaries).
So wouldn't you want to also capture information that indicates, say, *metaphorical* usage? For example, "die Tote Hose", (dee TO-tah HO-sah) in German might be accurately rendered in the New York City dialect of American English as "Fuhgeddaboudit!" [It means -- literally --"the dead trousers" and -- metaphorically -- "old news", "not worth talking about", etc.] This indicates the necessity for some level of meta-information, which is precisely what the Semantic Web is all about.
It seems like this could benefit from a Semantic Web interface of some sort. As other posters have noted, capturing contextual information is vital to adequate translation.
Perhaps this Semantic Web interface could be a third component, somewhere between the first SOAP protocol and the second SETI-like protocol, designed to give volunteers some kind of contextual clues to increase the accuracy of their translation.
BTW, some posters have also raised the question of "Trolls". Perhaps this could be avoided by first asking volunteers to rate the accuracy of other volunteers' translations. Maybe having a high meta-mod score would lead to increased "first translation" opportunities and decreased "this must be checked" translations.
DNA is a Turing machine. You, however, being dynamic and emergent, are not.