| Pages in topic: < [1 2 3 4] | Poll: Do you think that machine translation will significantly reduce the need for human translation? Thread poster: ProZ.com Staff
|
|---|
| No - and I speak from experience | Aug 9, 2012 |
I have worked with MT for 30 years, and I'm aware of the uses to which it is being put. It is most widely used in localization, an area of translation that wouldn't have flourished without the availability of MT, and intelligence searching. It is rarely used for general-purpose translation, and most institutions that have tried it for this type of application have given up.
The output of some systems is now quite respectable--occasionally. I can show you some recent output that is v... See more I have worked with MT for 30 years, and I'm aware of the uses to which it is being put. It is most widely used in localization, an area of translation that wouldn't have flourished without the availability of MT, and intelligence searching. It is rarely used for general-purpose translation, and most institutions that have tried it for this type of application have given up.
The output of some systems is now quite respectable--occasionally. I can show you some recent output that is very good. But even so, there is the cost of postediting, training posteditors, etc., etc., which managers rarely want to invest in.
For more on this subject, see articles on my website, http://www.murieltranslations.com/linguistics_mt_articles.html. They are still valid; little has changed. ▲ Collapse
| | | |
Simon Bruni wrote: Sophie Dzhygir wrote:
And I'm not talking about "automatic translation facilities on the Internet", Teresa, and not either about "using the crude output of MT systems"... Is there some other form of MT that I don't know about?
What you can find over the Internet is crap, we all agree about this. But large companies like Microsoft have their own sophisticated MT systems, with their own (previous, human) translated corpuses ans their own algorithms. The output of these systems in technical fields, in a pair like EN-FR, is impressive.
It's not about "using crude output" because the result needs to be edited. This is called post-edition. But as said above, the output is really good. I mean, sometimes it is fully messed up, but most of the time there is very little to edit, and sometimes there is nothing to edit at all. So all in all, it is very quick to do, if I remember well, you can post-edit at least 10.000 words a day, which is even quicker than editing human translation.
Let's do the maths: for a 10.000 words project:
- regular translation: 5 days translation + at least one day editing/profreading (depending on quality and workflow)
- post-editing : 1 day
So the work volume is cut by at least 6 for the translator.
Yes, that's what I call a "significant reduction"!
And this not just an imaginary situation, this is based on personal experience. I did post-editing 2-3 times five years ago. I wanted to know what it is, how it works, what's the output. I've seen. I know. And I quickly stopped doing it because that's shooting yourself in the foot.
On the one side, I wish no one would do it so that post-editing does not take over in the end. On the other side, I wish everyone in this conversation would do it at least once so that they could really know what we're talking about.
Oh and, I know your objection: it doesn't apply to my pair, it doesn't apply to my specialty field, etc. Sure. Agree. Not yet. 
| | | | | Not in the next 20 years | Aug 9, 2012 |
For people that want to find out what a text is about without having to pay for it, free MT tools are a treasure. But use of MT for professional translation? Not in the next 20 years.
I recall that back in 1982 our company started looking at Systran, to replace all translators. One engineer actually told me that we was so sorry for me, because I would be fired if the software was bought. Well, that was 30 years ago! And machine translation, though it has improved a lot, it still not... See more For people that want to find out what a text is about without having to pay for it, free MT tools are a treasure. But use of MT for professional translation? Not in the next 20 years.
I recall that back in 1982 our company started looking at Systran, to replace all translators. One engineer actually told me that we was so sorry for me, because I would be fired if the software was bought. Well, that was 30 years ago! And machine translation, though it has improved a lot, it still not ready for prime time. There is a very nice article about this at New America ( http://www.newamerica.net/publications/articles/2012/why_computers_still_can_t_translate_languages_automatically_67447 ).
My guess it that it will still take at least 20-30 years, but I would not be surprised if it takes 50. MT, like AI (Artificial intelligence), is an elusive target. And it's possible that certain kinds of texts (literature, poetry, law) will never be properly translated by machines.
In the meantime, if you want a good laugh about this issue, have a look at http://www.seo-translator.com/machine-translation-can-you-do-better-than-free/ ▲ Collapse
| | | | palilula (X)
United States
Local time: 03:05
German to English
+ ...
| I Am A Traditional Translator | Dec 1, 2013 |
I started as in-house interpreter and then as freelance translator just because a regular salary has never paid enough for a single person to make both ends meet.
As a valued technical translator, I was never short of assignments; I was never asked to buy "membership” and pay fees in order to get any assignment.
By being a seasoned translator, I never cared about CAT. I am observing how pages on Internet nowadays sound really funny due to using such tools, which lack of logic and... See more I started as in-house interpreter and then as freelance translator just because a regular salary has never paid enough for a single person to make both ends meet.
As a valued technical translator, I was never short of assignments; I was never asked to buy "membership” and pay fees in order to get any assignment.
By being a seasoned translator, I never cared about CAT. I am observing how pages on Internet nowadays sound really funny due to using such tools, which lack of logic and the common sense of human kind.
I think I will never go for CAT and never buy CAT software unless a company asks me to use (professionally in a common sense manner) such tools and provides the CAT software for free to me. I am not worried about being out of job; I took two more majors, which would help me by knowing the terminology, and even by being employed in the area of my other majors. ▲ Collapse
| | | | | Pages in topic: < [1 2 3 4] | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » Poll: Do you think that machine translation will significantly reduce the need for human translation? | 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 
