[…] Language industry experts concurred with what Packer had to say next: “We believe, along with most of the research and academic community…that the current approach of statistical, phrase-based MT has kind of reached the end of its natural life.”
Although he concedes that statistical MT is capable of producing technically accurate translations, “they don’t sound like they came from a human. They’re not natural, they don’t flow well,” Packer pointed out.
Packer, whose two-year-old Language Technology team has been working on MT, speech recognition, and natural language understanding, said they are part of the applied machine learning team whose goal is to take AI (artificial intelligence) and apply it, at scale, to Facebook products. More.
See: Slator
Subscribe to the translation news daily digest here. See more translation news.
Comments about this article
Spain
Local time: 18:05
Member
English to French
40 years to come up with the finding that a machine cannot produce output that sounds human...They should have had a closer look at C3PO.
Looking forward to neural systems, artificial intelligence, quantum physics applied to brain behaviours and all that.
Philippe
Edit: These new techs will excel with new languages like Facebookese, Northern-Facebookese, Eastern-Urban-Facebookese... See more
40 years to come up with the finding that a machine cannot produce output that sounds human...They should have had a closer look at C3PO.
Looking forward to neural systems, artificial intelligence, quantum physics applied to brain behaviours and all that.
Philippe
Edit: These new techs will excel with new languages like Facebookese, Northern-Facebookese, Eastern-Urban-Facebookese and other dialects. The translation profession with standard languages is safe.
[Edited at 2016-06-10 12:47 GMT] ▲ Collapse
Bosnia and Herzegovina
Local time: 18:05
Member (2009)
English to Croatian
+ ...
Packer described Facebook language as “extremely informal. It’s full of slang, it’s very regional.” He said it is also laden with metaphors, idiomatic expressions, and is riddled with misspellings (most of them intentional). Additionally, as in the rest of the world, there is a marked difference in the way different age groups communicate on Facebook.
Yes, they are humans, not robots. Impressive!
“I... See more
Packer described Facebook language as “extremely informal. It’s full of slang, it’s very regional.” He said it is also laden with metaphors, idiomatic expressions, and is riddled with misspellings (most of them intentional). Additionally, as in the rest of the world, there is a marked difference in the way different age groups communicate on Facebook.
Yes, they are humans, not robots. Impressive!
“It’s great that I can find my dishwasher manual online so I can figure out how to get the lemon seeds out of the ‘spinny’ thing. [But] it turns out the language that’s in that dishwasher manual has very little to do with the language people are using to talk to each other on Facebook.”
Yes, I don't talk to my friends in instructional lines. Again, a very original conclusion.
They are driven more by business than they are by science which is why they will have hard time progressing with this. Creating a technology that will process and produce nuances or original language would equal creating a machine that can think and has an independent thinking process of its own - which will happen when exactly? ▲ Collapse
United States
Local time: 12:05
Spanish to English
+ ...
Google recognized this problem in 2014 (that's why they changed to a paid API system to slow this trend down because everyone was having their websites... See more
Google recognized this problem in 2014 (that's why they changed to a paid API system to slow this trend down because everyone was having their websites auto-translated for free):
"Google admits 'garbage in, garbage out' translation problem":
http://www.theregister.co.uk/2014/02/06/google_translate_issue/
Open problems in machine translation:
https://www.youtube.com/watch?v=6UVgFjJeFGY&feature=youtu.be&t=12m0s
Machine translation in the 1950s was only 5 years away:
https://www.youtube.com/watch?v=K-HfpsHPmvw
How Cold War Computers Translated Russian
Part 1: https://www.youtube.com/watch?v=5sLbWItc33I
Part 2: https://www.youtube.com/watch?v=2ac41CO7Nr0
[Edited at 2016-06-10 17:19 GMT] ▲ Collapse
United Kingdom
Local time: 17:05
Member (2008)
Italian to English
The problem with statistical MT is that it uses matches found from bilingual documents on the internet. However, the more MT-translated posts, sites and articles that are published, the more this interferes with the statistical matches - to the point where human translation solutions get downgraded in favor of the machine translations.
V. interesting - that hadn't occurred to me. As you say: at a certain point the algorhythm will start picking up the machine translations done *by itself* and will memorise them, gradually worsening (not improving) the quality of the translations !
To report site rules violations or get help, contact a site moderator:
You can also contact site staff by submitting a support request »
This discussion can also be accessed via the ProZ.com forum pages.