Comments on: Google Adds Roundtables to Its Video Library http://webworkerdaily.com/2008/10/07/google-adds-roundtables-to-its-video-library/ Rebooting the workforce Wed, 25 Nov 2009 17:43:24 +0000 http://wordpress.com/ hourly 1 By: Nitin http://webworkerdaily.com/2008/10/07/google-adds-roundtables-to-its-video-library/#comment-306257 Nitin Wed, 08 Oct 2008 01:27:18 +0000 http://webworkerdaily.com/?p=4191#comment-306257 I apologize for the huge number of typos in my earlier comment. I hope the crux of my post was not as murky as its composition :) I apologize for the huge number of typos in my earlier comment. I hope the crux of my post was not as murky as its composition :)

]]> By: Nitin http://webworkerdaily.com/2008/10/07/google-adds-roundtables-to-its-video-library/#comment-306256 Nitin Wed, 08 Oct 2008 01:24:49 +0000 http://webworkerdaily.com/?p=4191#comment-306256 Samuel, I am a Ph.D. student and my research is in the area of statistical machine translation (SMT). I know it sounds like the field of machine translation has remained stagnant over the last 10 years but take it from someone who is looking at the research side of this. While it may seem to end users that not much has changed in this area, the amount of research that's been done over the last 10 years is absolutely phonomenal and the current research systems are SO much better than. Google Translate is a somewhat good proxy for such improvements but there are systems being developed right now (thanks primarily to DARPA's GALE program) that can beat easily beat Google given the same amount of data. I know it seems frustrating that not many of these improvements make it over to the consumer side of the equation, but I am confident that it will soon get better. As I tell me students and colleagues, translation is a good example of a problem that's easy for a human unless the human is telling a machine how to do it. Samuel,

I am a Ph.D. student and my research is in the area of statistical machine translation (SMT). I know it sounds like the field of machine translation has remained stagnant over the last 10 years but take it from someone who is looking at the research side of this. While it may seem to end users that not much has changed in this area, the amount of research that’s been done over the last 10 years is absolutely phonomenal and the current research systems are SO much better than. Google Translate is a somewhat good proxy for such improvements but there are systems being developed right now (thanks primarily to DARPA’s GALE program) that can beat easily beat Google given the same amount of data.

I know it seems frustrating that not many of these improvements make it over to the consumer side of the equation, but I am confident that it will soon get better. As I tell me students and colleagues, translation is a good example of a problem that’s easy for a human unless the human is telling a machine how to do it.

]]>