I was just reading an interesting discussion over on LinkedIn about Google’s efforts to create a natural language translation tool. Here’s the initial question:
“How many of you think that Googles anouncement last week of developing a speech to speech translation technoloy in natrural language in 2-3 years is feasible??” –Atiq Rehman
Here are some of the highlights of the discussion:
“You can find those statements from the 1950s on.” – Cerstin Mahlow
“As long as the sources of error are as great as they are now, I have trouble thinking of many contexts where people would be willing to tolerate the flaws. Maybe chatters who are only looking for entertainment and have no bottom line regarding accuracy. In war / emergency contexts, perhaps. In business, I think the problems just about doom the effort, unless a cultural adjustment makes people value “meeting” someone in this way even when the comprehension is shaky.” –John Rehling
“Underestimating Google is often a mistake, but, as an example, could I remind readers of the 2006 announcement by IBM of ‘Real Time Translation Services'” –Eric Janke
This is a very interesting discussion, as I work in both the technology and linguistic arenas I can see a bit of both sides.
On the linguistic side it is very difficult to imagine that a machine can replace a human when thinking of translation. There many different nuances to every language and the same word can often mean many different things. Getting the correct tanslation can be a very intuitive process. Let alone dealing with accents, dialects etc.
On the technology side it is a great and exciting challenge to create a tool that can intelligently translate language in real time. Computers are becoming more and more intelligent, and today they can logically solve problems that we could dream of 10-15 years ago.
I don’t believe that real time translation will be perfected anytime soon. It can be functional for very basic speech but as John said it must be very rudimentary. Computers can use context, formulas or other logical processes, but they just don’t have the ability to intuitively understand what is being said. Plus, pronunciations, accents, slang make it difficult as well. The speech would have to be very formal and correct to even have a chance.
What do you think?
You can check out the discussion Googles Recent Anouncement on LinkedIn in The Language Technology Group
