This seems to be the new trend in language translation. Using your smart phone to translate words in front of you in real time. Pleco seems to be the first one to have done so with their [url=http://tw.forumosa.com/t/pleco-chinese-dictionary-camera-recognizer-iphone-app/59966/1 character translation iPhone app[/url].
Word Lens is doing something similar arguably with even better implementation on their iPhone app. The app is free, but dictionaries with different languages will be sold separately. Looks pretty cool!
Yes, the two really cool things are that it displays the altered text in the same font / background as the original, and that you can save the image to make it easier to refer to.
[quote=“ironlady”]Yeeeeeahhh…and in their own promotional video, they allow an ungrammatical, nonsensical Spanish sentence like “y lo va el otro direccion” to show?
“Lo” is an object pronoun, not a nominative pronoun; “direccion” is feminine, not masculine, and the preposition “en” is totally missing.
:no-no:[/quote]
This is the dawn for this tech. It has not yet had its morning coffee. Give it a few espressos and it will be at full song. The march of this sort of thing to mainstream, widespread usage is inevitable. As are real-time speech translators. The only question is not whether but when.
We are less than 2 1/2 years into the age of big-screen, camera-toting, consumer-grade, third-party-application-running, pocket-computers with enough horsepower to even do this scenario. July 10, 2008 to December, 2010. 30 months from nothing to here. Any decent adoption will lead not only to more mindshare entering this arena, but also to more crowdsourcing of translation improvements from humans. Definite positive feedback loop. So we can be sure the next 30 months will bring tremendous improvements.
It seriously makes me wonder the real benefit to (opportunity) cost considerations of people learning a second language at this point, especially since actual learning is now in a race with technology. I’m not going to quit learning Chinese, but I’ll second guess the why of it when the going gets rough.
I don’t doubt that it WILL be better, and probably fairly soon. What I do question is the sense of any organization that puts out a promotional video with THAT showing. I mean, does anyone there actually speak Spanish? Or think that anyone who does might see their video?
I don’t think the all-too-common machine translation errors will put many people off. It’s to be expected, and most people will be more than a little forgiving for on the spot assistance.
From the feedback on iTunes, I think the company made a mistake in making the barebones engine available for free rather than simply making it a straight, paid download. Goofy people. I bought the Pleco dictionary plug-in for $15. I would have LEAPT to pay $25 for the same app had it featured this interface/display.
I expect that, flaws and all, the company behind this just scared a dozen potential competitors out of the market, and sent serious competitors back to the drawing board.
I have to agree for once. Back when I was a poor student, I did a one-year elective in book publishing. We learned typesetting and estimating. Even though the big houses were already using computers to take paper pulp in at one end and churn out paperbacks at the other. Total waste of time. These things are getting better and better, and it won’t be long until professional translators will only be used for the very top-level stuff, I think. There will be a HUGE body of human translators who will be finding themselves obsolete before too long, I think.
I don’t know if any of you speak Spanish… but in my estimation, a native speaker of Spanish without any English background probably would not understand that sentence. It is essentially meaningless garbled words in a row.
Errors can be forgiven – otherwise no one would do simultaneous interpretation. But this truly goes beyond “errors” into the realm of “what were they thinking”? I mean, you can pick your examples when you’re producing a short demo video!
Maybe for English to Spanish and vice versa machine translations are getting better. The fact remain that machine translator is always based on a set of algorithms that are revised in order to improve accuracy. That means the machine doesn’t really know what the text means, all it sees are digital numbers and the translation program translates based on the algorithm set by programmers.
Which means without a true artificial intelligence that are as smart as a human being who are self aware, it is impossible to get any kind of machine translation that will sound coherent out of the box. MT is good for people who wants a general meaning of something or perhaps a translation of some words for the purpose of language learning, but translators must understand culture which is something a machine will never “understand”.
As long as you get something like “Stir fry cattle river with no result” from 干炒牛河 with machine translation, human translators are in no danger of losing their job. The only real problem for Chinese to English is companies not knowing the difference between quality (and how it will affect the way their customers perceive them).
I mean, there are still a need for editors, and software grammar checker is actually pretty good.
[quote=“Taiwan Luthiers”]Maybe for English to Spanish and vice versa machine translations are getting better. The fact remain that machine translator is always based on a set of algorithms that are revised in order to improve accuracy. That means the machine doesn’t really know what the text means, all it sees are digital numbers and the translation program translates based on the algorithm set by programmers.
Which means without a true artificial intelligence that are as smart as a human being who are self aware, it is impossible to get any kind of machine translation that will sound coherent out of the box. MT is good for people who wants a general meaning of something or perhaps a translation of some words for the purpose of language learning, but translators must understand culture which is something a machine will never “understand”.
As long as you get something like “Stir fry cattle river with no result” from 干炒牛河 with machine translation, human translators are in no danger of losing their job. The only real problem for Chinese to English is companies not knowing the difference between quality (and how it will affect the way their customers perceive them).
I mean, there are still a need for editors, and software grammar checker is actually pretty good.[/quote]
I don’t know…google translate gave me “fried beef” for 干炒牛河 and even read it to me in melliflous tones. The enemy is at our door.
the “provide a better translation” box in google translate = crowdsourcing the translations from humans. when they get enough matching submissions, the MT adapts. “artificial artificial intelligence”.
a while ago some folks started using mechanical turk mygengo.com that employs this method as a business.
second life itself now has universal translation projects going on, which are unbound by the constraints of the physical world (batteries, adoption of miniature computers, etc.) and thus can progress much faster as long as there is enough brain power being put to the task. wiki.secondlife.com/wiki/Universal_Translator
personally, i think the world will become a much smaller place when we can buy a pair of sunglasses with the reading built in, complete with a universal translator in ear. apart from the obvious positive effect on the travel industry, there are other scenarios that will be possible. paired with mobile, this would have a huge impact on “reverse outsourcing”: the ability to connect experts who speak a first-world language to people in third-world who need services, who have more mobile phone penetration than people in the first world, and may not speak the language of their service providers.
The problem with anything tagged crowdsourcing in the translation field is that there is no bar to entry, so anyone can weigh in with what they believe is the “best” translation.
The minute you get thousands of Chinese from the PRC (for example) influencing Chinese>English AI translation engines, the quality will probably end up being worse than if the AI algorithm did the job alone. :aiyo: I’m not saying there are not some competent people from there, but when the masses get turned loose and start having fun, the result is pretty much like the laughable and patently unreliable “KudoZ” system at proz.com. It’s more important in that system to be liked than to be accurate, but there’s not much difference between that (since it is a smaller, closed community in that case) and what would happen when, say, a huge number of non-native-speakers contributed “against” a smaller number of native speakers in the target language.
Don’t get me wrong – I think things are going in this direction. I just don’t think it’s going to be pretty for a really long time.
Unless there are some kind of credential system where only qualified people are allowed to translate. To be honest with you some of the credential rating systems out there (like Cloud Crowd) are rated by people who happens to have high “creditability score” who may or may not be qualified.
Furthermore if something like a book or a game needs to be translated then cloudsourcing is a terrible way to get things done because the content will not be coherent. It will be like translating with a dictionary. Those projects are often best done with a small team who communicates with each other. Sometimes I wonder what would one get from a cloud-sourced translation of a book when each translator only sees a few page of the text (out of thousands).
If anything cloudsourcing will weed out those cheap Taiwanese/Chinese translation agencies who demands too much for too little money, when they find that they’re unable to continue their business because the cheap/free translation they get from cloudsourcing isn’t acceptable anywhere else.
I disagree. I don’t see it as any different from the way I can make sense out of a garbled sentence in Chinglish. A native speaker reads the garble, and mentally corrects the erroneous portions, and supplies the missing components.
I disagree. I don’t see it as any different from the way I can make sense out of a garbled sentence in Chinglish. A native speaker reads the garble, and mentally corrects the erroneous portions, and supplies the missing components.[/quote]
Has anyone here edited machine translated documents? From what I heard it’s a real nightmare.
I disagree. I don’t see it as any different from the way I can make sense out of a garbled sentence in Chinglish. A native speaker reads the garble, and mentally corrects the erroneous portions, and supplies the missing components.[/quote]
Has anyone here edited machine translated documents? From what I heard it’s a real nightmare.[/quote]
As long as you have the original text, it’s not so bad. I’m doing a website from Chinese to English and I’m using Google and what little Chinese I know. I also have the people who wrote it here that I can ask questions to. The Google gives you an idea of what they are trying to say.
But this truly goes beyond “errors” into the realm of “what were they thinking”? I mean, you can pick your examples when you’re producing a short demo video!