| Pages sur ce sujet: < [1 2] |
| Utilisateur | Auteur du fil: Csaba Ban After machine translation, here comes machine interpreting! |
OlafK Royaume-Uni Local time: 16:55 allemand + ... |
I'd love to see the faces of people listening to speeches by say George Bush, John Prescott or Boris Johnson instantaneously "translated" into Japanese! | | | |
FarkasAndras Hongrie Local time: 17:55 anglais vers hongrois + ... |
Williamson wrote:
Some who are 25-30 years old should better start worrying.
If a career, freelance or otherwise as an interpreter spans 4 decades, where will they be in say 2050 at the current rate of technological evolution. I know where I will be in 2050 : dead and gone. No worries for me.
Make no mistake, if the current Mekka of interpreting (E.U.-institutions)can save costs by gradually replacing interpreters by machines (robots) which are (almost) as perfomant as humans, it will be done.
Not by sacking people, but by not replacing them by youngsters. That's the way social minded state owned companies do it. |
|
Unless they constrain speakers (demand clear pronounciation and very simple, short sentences with very clear wording and no figures of speech), I don't think machine interpreting can get close to humans. This is not really the usual case of tecnological development: this is not a matter of raw computing power. It can't be solved by brute force, it needs very smart algorithms. Perhaps smarter than a computer can ever be.
Then of course, who knows... Maybe the truly massive databases that will be created in the next few decades and the smart analysis tools that are being developed will allow a machine to truly "understand" texts or at least act like it. | | | |
Madeleine MacRae Klintebo Royaume-Uni Local time: 16:55
Membre suédois vers anglais + ... | | PR agency definitely got results | Nov 7 |
This is, at least, the third thread about this product at proz. | | | |
Pablo Grosschmid Espagne Local time: 17:55
 Membre (2003) anglais vers espagnol + ... | | More fuel into the fire ! | Nov 7 |
Forget, robotics, chess, rule-based and statistical MT algorithms, they have little to do with interpretation.
Kedves András:
Not only it is not beside the point, it is the only way to do it.
It is true that programmers approach problems differently than humans. That is the problem in this case.
¿Why do I think so?
Speech recognition and voice synthesizing are already here, and rapidly improving.
The engine in the middle can not and will not be any MT algorithm (they are useless for this task), but a model replicating what interpreters do every day.
Storage capacity and processing speed are skyrocketing and becoming cheaper and cheaper, and it is just a matter of reaching the cost/benefit threshold.
Dear Parrot:
You got it, interpretation does not function literally (no need).
Models proposed are certainly complex.
If you are interested in reading one more of the "Copy what happens in Nature" type, just contact me.
Best wishes to all !
[Edited at 2009-11-07 17:28 GMT] | | | |
FarkasAndras Hongrie Local time: 17:55 anglais vers hongrois + ... |
Pablo Grosschmid wrote:
The engine in the middle can not and will not be any MT algorithm (they are useless for this task), but a model replicating what interpreters do every day. |
|
What makes you think so? Why do you think spoken text is so different from written text that a "traditional" MT algorithm wouldn't be the way to go?
As to replicating human thought processes as complex as those related to language, I can only repeat that computers are awful at it and get much better results through an approach that builds on their own strengths.
There are scientists out there working on replicating the human brain itself. Given its almost unparallelled complexity, this may never succeed... if it does, the final product might function like a human brain and might be able to interpret like a human. But this is pure science fiction, and if someone were to do it, they will probably have better uses for it than replacing human interpreters... | | | |
Mohamed Mehenoun Algérie Local time: 17:55
 Membre (2008) anglais vers français + ... | | I don't see it badly... | Nov 7 |
Hello,
As I see it a development in Robotics is needed if men want to conquer space, sending an "intelligent" machine to space would solve many technical problems...
Furthermore, the automatic interpretation would be very useful in this aspect as they would be able to send people from different countries, and they would be able to communicate perfectly without the help of a third party...
And the applications to machien interpreting/translation could be tremendously useful in terms of cost and applications !
The matter is very serious and could be compared to an industrial revolution...
The problem isn't what nature does, it's how to use it to go farther...
Again that's my opinion...
Regards, | | | |
Williamson Royaume-Uni Local time: 16:55 flamand vers anglais + ... |
an interpreting machine/robot? | | | |
Pablo Grosschmid Espagne Local time: 17:55
 Membre (2003) anglais vers espagnol + ... | | Very probable, it may be just around the corner | Nov 8 |
But not via MT, simply because translation and interpretation are two different kettle of fish, as written text is very different from spontaneous speech.
There are out there many good translators who are poor interpreters (and viceversa).
No need either to wait millenia until scientists, maybe, "replicate the human brain".
Interpreting by humans works well enough, and it is a fairly straightforward mechanical task that can be done without having to "think".
Did you ever see someone in the booth interpreting from language A into language B, and at the same time reading a newspeper, writing a letter or solving crossword puzzles in a third language? I have seen it, more than once.
Interpreters only have to find in their hard disk the best matches for what is called "units of sense" (nothing to do with traditional grammar).
This should be child's play for a modern computer, provided it has been fed with the routines and data used by human interpreters, and refined by training. | | | |
VitoSmolej Allemagne Local time: 17:55
 Membre (2004) anglais vers slovène + ... | | one step further .... | Nov 9 |
Claudio Porcellana wrote:
but why stopping only here?
software companies can even create a MP (machine proofreader)...
|
|
... or a machine client, so that the machine agency can do without the human interface. | | | |
Claudio Porcellana Italie Local time: 17:55
 Membre (2004) anglais vers italien | | one step further .... | Nov 9 |
hàhàhà
great answer VitoSmolej

Claudio
P.S. clearly, this post comes from a machine teller ...
[Modificato alle 2009-11-10 11:06 GMT] | | | |
| Pages sur ce sujet: < [1 2] |