Google Begins Using High-Accuracy Neural Machine Translation System For Eight Language Pairs

Status
Not open for further replies.
the new Neural Machine Translation system is that it analyzes the sentence as a whole, rather than as the sum of its parts.
Only sentences? So, each sentence in a paragraph is translated apart from all the others? I guess it's a step in the right direction... It's explained here: https://www.youtube.com/watch?v=l2dVjADTEDU Worth watching, if only for his (albeit simplified) take on what a thought actually is.

Man, I hope they add dolphin and whale dialects, pretty soon.

Seriously, if there are any Douglas Adams fans out there, just think about how close we're coming to having social media arguments with all the non-English speakers of the world.
Meanwhile the poor Babel fish, by effectively removing all barriers to communication between different cultures and races, has caused more and bloodier wars than anything else in the history of creation.
https://en.wikipedia.org/wiki/List_of_races_and_species_in_The_Hitchhiker%27s_Guide_to_the_Galaxy#Babel_fish
 
Google Translate is getting really good. Now all that's left is to identify mistakes, so we can actually use it with confidence.
Google Translate still can't cope with common mistakes, like "Allot of things". It does ask "did you mean: a lot of things?", but we wouldn't be asking for a translation if we knew. It also has difficulties noticing the relationship between two phrases if separated by a dot, like "I don't know. Do you?".

But it's a step in the right direction. Thumbs up for Google!
 
impressive considering most humans have trouble translating between two languages.. in fact some, most notably American, have trouble with just one language.. LOL
 
Status
Not open for further replies.