Wow this is really interesting, do you guys see this eventually being a replacement for interpretors outside of military operations (i.e. U.N., international business etc.) cause I'm training to become an interpretor! Lol don't want to lose my job to a Nexus One!! haha
Hitch-Hiker's Guide to the Galaxy predicted this with the Babel-Fish! It's fascinating to see this technology being put to good use; this is an absolute necessity for the protection of the troops and civilians alike. I wonder: is this technology capable of interpreting variations in local dialects and offering 'sets' of possible interpretations to either/both end-users in situations where a translation is ambiguous, thereby replicating the more precise skills one would expect from a human?
From Craig Schlenoff, Project Leader, NIST TRANSTAC Evaluation Team
For this effort, we have focused on one dialect of Pashto, namely Kandahari. However, when the same systems were put in front of people who speak other dialects, they still worked, though not as well. The systems are only as good as what they are trained on … if you train them on other dialects, they should work equally well on them.
Thanks for checking out our videos. Please add your comments and let us know what you think. We will be reviewing and then posting comments as long as they are on topic, respectful and do not promote specific products or services.
From Craig Schlenoff, Project Leader, NIST TRANSTAC Evaluation Team
The systems tend to work better on shorter, simpler sentences, but can still be used for more complex sentences. From initial analysis, utterances that are less than 15 words seem to work best with roughly linear degradation after that as a function of the number of words.
@usnistgov I see. Does the time taken to translate a sentence increase depending on its length? For instance, how long would the previous sentence take to translate (roughly)?
From Craig Schlenoff, Project Leader, NIST TRANSTAC Evaluation Team
The time the systems takes to translate a sentence does vary with the length of the sentence. A few words sentence may only take a second or two while a longer sentences (10 words) may take 4-5 seconds.
wow Great.I lOVE Pashto.It is Great language.In Pashto say:Ta Sanga Ye ?it means how are you? Za Kha Yam Its means I am Fine :):):
powerpashtoons 1 year ago
That is really cool.
samh09 1 year ago
I want a klingon version...
applekidtw 1 year ago
Wow this is really interesting, do you guys see this eventually being a replacement for interpretors outside of military operations (i.e. U.N., international business etc.) cause I'm training to become an interpretor! Lol don't want to lose my job to a Nexus One!! haha
dagreatone2009 1 year ago
@dagreatone2009 I heard that Talibans put a higher price on capturing interpretors than US soldiers. Are you training for this dialect?
kalucardable 1 year ago
curiously enough that StarTrek translator device seems more like StarWars lightsaber(TM)...
captainFLUORESCENT 1 year ago
Hitch-Hiker's Guide to the Galaxy predicted this with the Babel-Fish! It's fascinating to see this technology being put to good use; this is an absolute necessity for the protection of the troops and civilians alike. I wonder: is this technology capable of interpreting variations in local dialects and offering 'sets' of possible interpretations to either/both end-users in situations where a translation is ambiguous, thereby replicating the more precise skills one would expect from a human?
balboarock 1 year ago
From Craig Schlenoff, Project Leader, NIST TRANSTAC Evaluation Team
For this effort, we have focused on one dialect of Pashto, namely Kandahari. However, when the same systems were put in front of people who speak other dialects, they still worked, though not as well. The systems are only as good as what they are trained on … if you train them on other dialects, they should work equally well on them.
usnistgov 1 year ago
That was the nexus one... anyway..
conorsmurf 1 year ago
Thanks for checking out our videos. Please add your comments and let us know what you think. We will be reviewing and then posting comments as long as they are on topic, respectful and do not promote specific products or services.
usnistgov 1 year ago
@usnistgov Just curious, how complex a sentence can these devices translate?
AndrewFaulds 1 year ago
From Craig Schlenoff, Project Leader, NIST TRANSTAC Evaluation Team
The systems tend to work better on shorter, simpler sentences, but can still be used for more complex sentences. From initial analysis, utterances that are less than 15 words seem to work best with roughly linear degradation after that as a function of the number of words.
usnistgov 1 year ago
@usnistgov I see. Does the time taken to translate a sentence increase depending on its length? For instance, how long would the previous sentence take to translate (roughly)?
AndrewFaulds 1 year ago
From Craig Schlenoff, Project Leader, NIST TRANSTAC Evaluation Team
The time the systems takes to translate a sentence does vary with the length of the sentence. A few words sentence may only take a second or two while a longer sentences (10 words) may take 4-5 seconds.
usnistgov 1 year ago