| Pages in topic: < [1 2] | "Stanford system combines software with human intelligence to improve translation" - article Thread poster: Astrid_H
|
|---|
| To me, it sounds just like ... | Nov 3, 2014 |
Meta Arkadia wrote:
it's not Auto-Suggest, Auto-Write, Auto-Complete, or whatever the CAT tool vendors call it. It's more like Auto-Assemble as available in DejaVu and CafeTran, but on a much higher level. I think it's frightening.
... what Atril Déjà Vu X3's "AutoWrite" already does when coupled to an MT system that can "learn" from the translator, like PROMPT.
What's so frightening about it?
| | | |
David Turner wrote:
... what Atril Déjà Vu X3's "AutoWrite" already does when coupled to an MT system that can "learn" from the translator, like PROMPT.
As I said, it is NOT what DV and CafeTran already do. It is NOT AutoWhatever. Did you read the article?
AutoWhatever (here "source word lookup)requires user/translator activity: You'll have to typ n letters before you'll get a suggestion. It's relatively easy, a regex will do.
Target Autocomplete just shows up, and if you look at the picture, the suggestions show up in the right place of the sentence.
Cheers,
Hans
[Edited at 2014-11-03 11:36 GMT]
| | | | | Just because "you" said it... | Nov 3, 2014 |
As I said, it is NOT what DV and CafeTran already do
... it doesn't mean that it's necessarily true....
Didn't have time to read the whole article, just a couple of paragraphs on how it works. But to me, it looks more like DVX3's AutoWrite than AutoAssemble.
In the example, the translator has typed "Teachers", so this AutoWhatever does seem to work on the basis of what the translator enters. If you don't enter anything, you get the MT for the whole sentence. If you enter the first word, it recalculates the MT for the remainder of the sentence, which is quite clever, but would translators really want to work this way?
It's true that DVX3's AutoWrite does need at least "one" letter to be entered but from then on it's quite capable of completing whole sentences with just a couple of keystrokes using your Lexicon, TB, TM and DeepMiner (as well as MT if it doesn't find what it wants). I think it's a more natural process for autocompletion to be closely geared to what the translator is typing and AutoWrite does this quite successfully as far as I'm concerned.
| | | |
David Turner wrote: Just because "you" said it... it doesn't mean that it's necessarily true....
It is.
Cheers,
Hans
[Edited at 2014-11-03 14:33 GMT]
| | |
|
|
|
| The Translator | Nov 3, 2014 |
David Turner wrote:
I think it's a more natural process for autocompletion to be closely geared to what the translator is typing and AutoWrite does this quite successfully as far as I'm concerned.
You miss the point entirely. This isn't a CAT tool. We can't buy or use it. It's no competition to Atril. You don't have to worry, but if you want to worry, worry about CafeTran.
MT doesn't care about translators. They are to be eliminated. Well, they can live a little longer to teach the MT how to function, and they'll get water and bread for the time being, but that's it.
Cheers,
Hans
[Edited at 2014-11-03 14:15 GMT]
[Edited at 2014-11-03 14:36 GMT]
| | | |
You miss the point entirely.
Sorry about that then.
It is a "system" though and I don't see why DVX's AutoWrite or Cafetran's autowhatever couldn't be upgraded, if it actually is an improvement, to emulate it.
As far as worrying is concerned, I sometimes worry about my inability to see what's so great about the Cafetran you and Michael so unstintling bang the drum for. Old dogs can't learn new tricks I guess.
| | | | John Fossey 
Canada
Local time: 19:51
Member (2008)
French to English
+ ...
| Effects of new technology | Nov 3, 2014 |
Astrid_H wrote:
I'm always curious about these things, so after thinking about it for a while, I don't think this new development is not necessarily just scary or dangerous, but it will quite certainly bring change to the business, if it is as good as it sounds. Maybe like horse carts being complemented by cars and then later replaced by them?
Well, did the newfangled cars put the transportation companies out of business? Some that couldn't keep up with change undoubtedly insisted on their horses and went out of business, but those who kept up with the new technology were able to contribute to a vastly higher volume of business than was previously able to be done.
There's a lot left to be translated yet.
| | | | Astrid_H
Germany
Local time: 01:51
German to English
+ ...
TOPIC STARTER
John Fossey wrote: Astrid_H wrote:
I'm always curious about these things, so after thinking about it for a while, I don't think this new development is not necessarily just scary or dangerous, but it will quite certainly bring change to the business, if it is as good as it sounds. Maybe like horse carts being complemented by cars and then later replaced by them? Well, did the newfangled cars put the transportation companies out of business? Some that couldn't keep up with change undoubtedly insisted on their horses and went out of business, but those who kept up with the new technology were able to contribute to a vastly higher volume of business than was previously able to be done. There's a lot left to be translated yet.
...which is exactly what I meant to express .
| | |
|
|
|
John Fossey wrote:
Well, did the newfangled cars put the transportation companies out of business?
Not the transportation companies, most of them. But the horses...
Cheers,
Hans
| | | |
David Turner wrote:
It is a "system" though and I don't see why DVX's AutoWrite or Cafetran's autowhatever couldn't be upgraded, if it actually is an improvement, to emulate it.
Which is why you should read the article. Atril was the first to implement Auto-Assemble - an MT technology - in a CAT tool, and we benefited immensely from it. It's not impossible that this part of the MT technology can also be ported to CAT tools.
As far as worrying is concerned, I sometimes worry about my inability to see what's so great about the Cafetran you and Michael so unstintling bang the drum for. Old dogs can't learn new tricks I guess.
You're probably way older than I am, yes. Most people over 61 are. All of them, actually. However, you apparently judge without reading articles, and without trying CAT tools. I'll show you one reason to try CafeTran, a screencast. Don't worry, it's only about two minutes in which I perform 4 searches in the 2 million segments DGT database, two searches in that DGT as a TMX memory, two as an indexed H2 database. Those results alone should make you try CafeTran.
Cheers,
Hans
| | | | | It seems to me that you're the one doing the "judging" Hans... | Nov 4, 2014 |
However, you apparently judge without reading articles, and without trying CAT tools.
... And a bit aggressively, I might add. I've just been giving my humble opinion.
As I said, in one of my replies, I did read the relevant part of the article (I've read the whole thing now but that hasn't changed anything).
I've also tried a good many CAT tools, or at least given them a brief test drive, all the way back to Optimizer, IBM Translation Manager, TWB 1.1, etc. in the early 1990s.
As for Cafetran, I've downloaded it 5 or 6 times, read the getting started instructions, read your own blog, and followed the Google group, to some extent, so please don't accuse me of not trying.
Anyway, enough said. This is way off-topic.
| | | | | It looks slightly more sophisticated than an auto-suggest-type function... | Nov 4, 2014 |
...in terms of the quality of suggestions it gives you...but not much.
They say it "adapts" to your "translation style" (translation: it keeps track of what you've entered in the past), but there's no "supercalifragilistic" AI working in the background.
The improvement, if any, appears to be quantitative in nature (i.e. helps you translate more quickly), and not qualitative (it won't turn you into an "Henri Bergson of translation" by any means).
My 2 cents... See more ...in terms of the quality of suggestions it gives you...but not much.
They say it "adapts" to your "translation style" (translation: it keeps track of what you've entered in the past), but there's no "supercalifragilistic" AI working in the background.
The improvement, if any, appears to be quantitative in nature (i.e. helps you translate more quickly), and not qualitative (it won't turn you into an "Henri Bergson of translation" by any means).
My 2 cents.
Michael ▲ Collapse
| | | | | Pages in topic: < [1 2] | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » "Stanford system combines software with human intelligence to improve translation" - article | Wordfast Pro | Translation Memory Software for Any Platform
Exclusive discount for ProZ.com users!
Save over 13% when purchasing Wordfast Pro through ProZ.com. Wordfast is the world's #1 provider of platform-independent Translation Memory software. Consistently ranked the most user-friendly and highest value
Buy now! » |
| | Anycount & Translation Office 3000 | Translation Office 3000
Translation Office 3000 is an advanced accounting tool for freelance translators and small agencies. TO3000 easily and seamlessly integrates with the business life of professional freelance translators.
More info » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |
Login to reply/comment 
Translation news
