| Pages in topic: [1 2] > |
Simultaneous translation by computer is getting closer
|
|---|
| |
| I love that cartoon | Jan 4, 2013 |
which basically says “will international relations get worse if this technology succeeds? Wil the earth be annihilated?”
Other problems aside, I find this paragraph the most amusing:
“One big difficulty when translating conversations is determining who is speaking at any moment. Mr Powell’s system does this not by attempting to recognise voices directly, but rather by running all the speech it hears through two translation engines simultaneously: English to Spanish,... See more which basically says “will international relations get worse if this technology succeeds? Wil the earth be annihilated?”
Other problems aside, I find this paragraph the most amusing:
“One big difficulty when translating conversations is determining who is speaking at any moment. Mr Powell’s system does this not by attempting to recognise voices directly, but rather by running all the speech it hears through two translation engines simultaneously: English to Spanish, and Spanish to English. Since only one of the outputs is likely to make any sense, the system can thus decide who is speaking. That done, it displays the translation in the other person’s goggles.”
This means this technology will not work in places where speakers routinely mix words or even sentences from more than one language. ▲ Collapse
| | | |
| Yes, I believe that... | Jan 4, 2013 |
... but I don't believe to the existence of "slowly speaking Spanish speakers"!
| | | |
Tom in London
United Kingdom
Local time: 23:28
Member (2008)
Italian to English
Simultaneous translation by computer is getting closer, in the same sense in which the earth is getting closer to the sun.
[Edited at 2013-01-04 21:17 GMT]
| | | |
felicij 
Local time: 00:28
German to Slovenian
+ ...
| And the computer | Jan 4, 2013 |
will be able to recognize accents, dialects and slangs. I haven't yet met a person that speaks perfect literary language...
And what about a stutterer?
Like Tom said, we will not live to see it substitute a real person.
| | |
|
|
|
Let's see first a computer getting the usual, text translation right
| | | |
| First things first | Jan 6, 2013 |
Miroslav Jeftic wrote: Let's see first a computer getting the usual, text translation right
No, I think the approach is right. Spoken language first. Linguistically, written language doesn't count. I even think that "converting" written text to spoken text before the translation followed by the reverse process would yield better results for documents. Can I patent this?
Cheers,
Hans (who strongly believes in MT)
| | | |
| |
| First things first | Jan 7, 2013 |
Meta Arkadia wrote:
No, I think the approach is right. Spoken language first. Linguistically, written language doesn't count. I even think that "converting" written text to spoken text before the translation followed by the reverse process would yield better results for documents. Can I patent this?
Why would that be the case? Computationally, spoken speech is much harder than written language, so if we got the harder problem solved, of course the easier problem should be already solved, shouldn’t it?
And spoken speech might not be a problem for German or English, but just imagine the number of homophones you will (not can) run into when you deal with spoken Chinese.
| | |
|
|
|
ATIL KAYHAN
Türkiye
Local time: 01:28
Member (2007)
Turkish to English
+ ...
| At the same time, | Jan 7, 2013 |
Chaos by computer is getting closer.
| | | |
Nigel Greenwood (X)
Spain
Local time: 00:28
Spanish to English
+ ...
| I'd love to see that... | Jan 10, 2013 |
however, I coincide with some of my colleagues (the majority) in that computer technology cannot recognise all the small -but important- deviations in spoken language that are used by everyone. I mean the use of coined expressions, which do not even exist in dictionaries. I work as a simultaneous/consecutive interpreter for aviation related training courses. I can assure you, that with all the "small" yet significant expressions used in this field, no computer will be able to do the job.
<... See more however, I coincide with some of my colleagues (the majority) in that computer technology cannot recognise all the small -but important- deviations in spoken language that are used by everyone. I mean the use of coined expressions, which do not even exist in dictionaries. I work as a simultaneous/consecutive interpreter for aviation related training courses. I can assure you, that with all the "small" yet significant expressions used in this field, no computer will be able to do the job.
Regards to all,
Nigel. ▲ Collapse
| | | |
| Love that video, Siegfried. | Jan 10, 2013 |
Thank you so much.
Gudrun
| | | |
Kaspars Melkis
United Kingdom
Local time: 23:28
English to Latvian
+ ...
In ironic twist, Twitter uses humans in its "computation engine".
They pay peanuts, but it can work for Twitter as there is nothing critical going on. We are not talking about international commerce or law here.
| | |
|
|
|
| First things first | Jan 10, 2013 |
Meta Arkadia wrote: Miroslav Jeftic wrote: Let's see first a computer getting the usual, text translation right No, I think the approach is right. Spoken language first. Linguistically, written language doesn't count. I even think that "converting" written text to spoken text before the translation followed by the reverse process would yield better results for documents. Can I patent this? Cheers, Hans (who strongly believes in MT)
You can't patent this. I have realised it's the only way back in the nineties. I have seen a presentation on Google Tech Talks showing same idea. Translating phonemes instead of text simplifies grammar parsing a lot.
| | | |
Jacek Podkanski wrote:
I have realised it's the only way back in the nineties.
I got the idea when translating Kurzweil's The Age of Virtual Machines into Dutch in 2000. I mentioned it a couple of times on translator forums, but it was always met strong unbelief. Your response is the first positive one. On the concept, that is.
It's increasingly frustrating to talk about MT with colleagues anyway. Their denial of MT is probably ostrich policy. MT is going to stay, and it will get better. Fast. The only way to survive as a translator is to be a darned good writer. And then you'll still lose.
Cheers,
Hans
| | | |
| |
| Pages in topic: [1 2] > |
Login to reply/comment 
Translation news
