| Pages in topic: [1 2] > | Poll: Who is more knowledgeable on machine translation and related technology? Thread poster: ProZ.com Staff
|
|---|
This forum topic is for the discussion of the poll question "Who is more knowledgeable on machine translation and related technology?".
View the poll results »
| | | | neilmac
Spain
Local time: 07:40
Spanish to English
+ ...
I don't often work with agencies, and I assume that most of my clients know little about the ins and outs of MT.
| | | |
I don't know, I’ve never discussed it with my clients. What’s this obsession with MT?
P.S. Some of my clients have added to their POs or NDAs a note saying that I should not use MT under any circumstances.
| | | | Mr. Satan (X)
English to Indonesian
| Q. Who is more knowledgeable on machine translation and related technology? | Sep 17, 2022 |
A. Computer scientists.
| | |
|
|
|
Sadek_A 
Local time: 09:40
English to Arabic
+ ...
| MT this, MT that | Sep 17, 2022 |
Teresa Borges wrote:
What’s this obsession with MT?
Showcase it heavily in the hope that everyone will surrender and submit to it!
| | | | William Hepner
United States
Local time: 01:40
Member (2022)
Spanish to English
| In the academic world, machine translation is a subfield of computational linguistics | Sep 17, 2022 |
So I'd have to say that either a computational linguist, a computer scientist/programmer, or someone in a related field would know more about the process and algorithms than, say, a translator (that is, an otherwise untrained translator).
To me, this seems like a strange question; usually this kind of "who is more..." interrogative sentence gives us at least two options to decide between. In any case, I'm not sure exactly who would know the most about machine translation, since it'... See more So I'd have to say that either a computational linguist, a computer scientist/programmer, or someone in a related field would know more about the process and algorithms than, say, a translator (that is, an otherwise untrained translator).
To me, this seems like a strange question; usually this kind of "who is more..." interrogative sentence gives us at least two options to decide between. In any case, I'm not sure exactly who would know the most about machine translation, since it's pretty vague as a question. The above is my best guess. ▲ Collapse
| | | |
It's likely that some people among my clients know more about MT than I do. It's also likely that some people among my clients know less.
It's not something we ever discuss. And like others who have posted, I wonder why we keep getting prodded into discussing it here.
To me, MT is just another tool, one that can be helpful and useful in some contexts, but not in others. Like spell-checkers and search and replace. Yet we aren't constantly being invited to discuss the mer... See more It's likely that some people among my clients know more about MT than I do. It's also likely that some people among my clients know less.
It's not something we ever discuss. And like others who have posted, I wonder why we keep getting prodded into discussing it here.
To me, MT is just another tool, one that can be helpful and useful in some contexts, but not in others. Like spell-checkers and search and replace. Yet we aren't constantly being invited to discuss the merits and failings of spell-checkers and search and replace on these forums. ▲ Collapse
| | | | Lieven Malaise
Belgium
Local time: 07:40
Member (2020)
French to Dutch
+ ...
| Definitely me. | Sep 18, 2022 |
If I'm offered MTPE jobs it's always in the shape of a pre-translated SDL Studio file. One of my clients uses DeepL, so the output is rather good, another uses something else and the output definitely isn't that good. But it doesn't matter, because I have developed my own way of treating MTPE projects in which I am able to translate faster and at a my own high quality level (no, not by treating it as a 'usual' translation ).
... See more If I'm offered MTPE jobs it's always in the shape of a pre-translated SDL Studio file. One of my clients uses DeepL, so the output is rather good, another uses something else and the output definitely isn't that good. But it doesn't matter, because I have developed my own way of treating MTPE projects in which I am able to translate faster and at a my own high quality level (no, not by treating it as a 'usual' translation ).
The fact that a lot of agencies use MTPE mainly to be able to spend less mony, also means they are not broadly going to invest in 'research' to explore every single possibility of the technology (that costs time and money!). That benefits translators taking their time and looking for ways to get the most out of it, while clients don't bother to do more than a quick MT pre-translation and let you cope with the rest.
Meanwhile translation represents still 80% of what I do (project-wise, that is, word-wise MTPE wouldn't even represent 3% of what I do). But I sure as hell will be ready for a less pure translation-oriented future. ▲ Collapse
| | |
|
|
|
expressisverbis
Portugal
Local time: 06:40
Member (2015)
English to Portuguese
+ ...
Generally speaking, I think it's the translators, because no one knows better than users what their needs are.
The vast majority of software is developed without taking into account the real needs of its users and... that's why sometimes we come across technical issues that should be considered before these programs are released.
Why not discuss terminology extraction tools, for a change?
| | | |
expressisverbis wrote:
Generally speaking, I think it's the translators, because no one knows better than users what their needs are.
The vast majority of software is developed without taking into account the real needs of its users
Do you really think the producers of MT see translators as their user base?
I don't. I think they see MT as something that will eventually replace translators, so the users are other people, either translation and localisation agencies, or just businesses that want to automate the entire translation process for their product documentation. So I don't expect MT developers to consult translators.
I think most people reading these forums don't expect to lose their jobs to an artificial intelligence any time soon, for reasons that have already been discussed to death, but it seems likely to me that a lot of mass translation of standard items - the kind of source material that benefits most from the use of CAT tools - will be dealt with automatically before too long.
I would also expect there to be some kind of battle between big agencies that have developed their own in-house translation engines and want to convince clients that theirs is the best. But that will be a temporary, intermediate stage, until we reach the point where a multilingual translation service is a built-in feature in Windows 15.
But I don't expect the need for precise, consistent translation of either critical technical material or literature, carried out by human specialists, to disappear within my own lifetime.
| | | | John Silva
Brazil
Local time: 02:40
English to Portuguese
+ ...
If we are talking about agencies, they probably know more. On the other hand, direct clients usually don't have this kind of knowleged. I have alreday been in situations where I had to explain a lot of the translation process to a client before working.
| | | | Kay Denney
France
Local time: 07:40
French to English
I expect my agency clients know more than me, since I don't use it and some of them do. And I expect my direct clients know less than me, since they've not looked into it at all.
A direct client asked me about it once, apparently a translator friend of hers told her it would be useful. I explained that no, it wouldn't be useful, because her content was creative and MT was only good if you were trying to say things that had already been said by someone else, whereas her work had to ... See more I expect my agency clients know more than me, since I don't use it and some of them do. And I expect my direct clients know less than me, since they've not looked into it at all.
A direct client asked me about it once, apparently a translator friend of hers told her it would be useful. I explained that no, it wouldn't be useful, because her content was creative and MT was only good if you were trying to say things that had already been said by someone else, whereas her work had to be original. She was apparently relieved with my answer. ▲ Collapse
| | |
|
|
|
Mihai Badea (X)
Luxembourg
Local time: 07:40
| Unknown asker | Sep 19, 2022 |
It is a shame we don't know the author of the question. We could have asked him/her: what's the assumption here? If an agency says it's near-human quality, the translator should accept an editing rate, without checking?
It is always a good idea to check. Perhaps the client sent the wrong file.
| | | | expressisverbis
Portugal
Local time: 06:40
Member (2015)
English to Portuguese
+ ...
| Users and software in general | Sep 19, 2022 |
Philip Lees wrote:
Do you really think the producers of MT see translators as their user base?
I don't. I think they see MT as something that will eventually replace translators, so the users are other people, either translation and localisation agencies, or just businesses that want to automate the entire translation process for their product documentation. So I don't expect MT developers to consult translators.
I think most people reading these forums don't expect to lose their jobs to an artificial intelligence any time soon, for reasons that have already been discussed to death, but it seems likely to me that a lot of mass translation of standard items - the kind of source material that benefits most from the use of CAT tools - will be dealt with automatically before too long.
I would also expect there to be some kind of battle between big agencies that have developed their own in-house translation engines and want to convince clients that theirs is the best. But that will be a temporary, intermediate stage, until we reach the point where a multilingual translation service is a built-in feature in Windows 15.
But I don't expect the need for precise, consistent translation of either critical technical material or literature, carried out by human specialists, to disappear within my own lifetime.
I was talking about software in general and not specifically about MT, such as accounting software, assisted translation software, project management software, etc.
However, if MT is made available to translators and agencies, then we become its users (of course it only applies to those who use it).
There has already been much debate on this issue. Human translations can't be replaced because translators use methods to approach/interact with each other, and machines cannot do this. I mean MT cannot interpret emotions and it never will.
However, it's useful to accelerate procedures and assist human translation in some scenarios. I see MT as just another tool in translation. We can love it or hate it.
I don't think that will happen. The use of MT is still very risky no matter how good it is or not, and no company would like to see its brand ruined.
Also, translators wouldn't like to see their reputation ruined.
There are more clients aware of this danger, than clients who adhere to MT. And the same goes to any decent and wise translator.
Let's hope not
Neither do I. In this respect, I consider myself a more optimistic person.
| | | | Mr. Satan (X)
English to Indonesian
| Anonymous poll | Sep 19, 2022 |
Mihai Badea wrote:
It is a shame we don't know the author of the question. We could have asked him/her: what's the assumption here?
When the poll question was submitted by a member, usually they would be credited and have their name mentioned in the original post. When the poll was anonymous, that means the question was created by a site staff.
Also, welcome back, Mihai. How was your vacation? 
| | | | | 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 » Poll: Who is more knowledgeable on machine translation and related technology? | 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 » |
| | CafeTran Espresso | You've never met a CAT tool this clever!
Translate faster & easier, using a sophisticated CAT tool built by a translator / developer.
Accept jobs from clients who use Trados, MemoQ, Wordfast & major CAT tools.
Download and start using CafeTran Espresso -- for free
Buy now! » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |
Login to reply/comment 
