Saturday, January 14, 2017

Camellia japonica: How best to add the remaining 6,994 known living languages, Jeff - https://scott-macleod.blogspot.com/2017/01/zebu-cows-breakthrough-googles-zero.html - to the 103 languages in Google Translate with the new GNMT, - and as what I'm calling as a new method "ethno-wiki-virtual-world-graphy"?, Jeff Dean via Google+ - "Zero-shot translations: translate Korean->Japanese without ever seeing Korean->Japanese training data," re the WUaS Universal Translator with Google Translate / Zero Shot?, Stanford Law Codex presentation and Language / Translation strategies?, Harvard Law JuryX course with Professor Charles Nesson, Stanford Law China Guiding Cases Project, In what ways could World University bring together legal translation with Stanford Law Codex re a high bar for generating excellent text versions in other languages, and from HarvardX video too?


In a thread with Jeff Dean's main post, who is a senior engineer at Google, and whom I've met at Stanford, I asked the following question here - https://research.googleblog.com/2016/11/zero-shot-translation-with-googles.html - re the World University and School Universal Translator with Google Translate / Zero Shot http://worlduniversity.wikia.com/wiki/WUaS_Universal_Translator ...

How best to add the remaining 6,994 known living languages, Jeff - https://scott-macleod.blogspot.com/2017/01/zebu-cows-breakthrough-googles-zero.html - and as what I'm calling as a new method "ethno-wiki-virtual-world-graphy" ?


*

https://plus.google.com/+JeffDean/posts/Q4XSDrPpfLF






Jeff Dean via Google+

1 month ago  -  Shared publicly
Zero-shot translations: translate Korean->Japanese without ever seeing Korean->Japanese training data.

One of the very interesting aspects of our recent work on neural machine translation is that we can train a single, multi-lingual model that is able to translate between many language pairs (say, English->Japanese, English->Korean, Japanese->English, and Korean->English). This often is able to improve quality for many of the language pairs, and also simplifies many production aspects of the system. However, a really interesting property is that a system trained on these language pairs is actually able to do a reasonable job of translating between language pairs where it has never seen training data before (say, Korean->Japanese).

This blog post gives the high level details:

https://research.googleblog.com/2016/11/zero-shot-translation-with-googles.html

Our Arxiv paper has details about this approach:

https://arxiv.org/abs/1611.04558

*

Zebu cows: Breakthrough > Google's Zero-Shot Translation & new LANGUAGE Inter-lingua > How to add the remaining 7,097 known living languages minus Google's 103 languages = 6,994 languages ~ to the WUaS_Universal_Translator, "Zero-Shot Translation with Google’s Multilingual Neural Machine Translation System," (From the TV show from the 1970s' "Mission Impossible" - "Your job should you chose to accept it ... " :) And how best to build this in to developing World University newly in Wikidata / Wikiversity beginning with Wikipedia's 358 languages?, It's kind of a question too for what I'm calling ethno-wiki-virtual-world-graphy and re a STEM centric realistic virtual earth and as class room too ...

https://scott-macleod.blogspot.com/2017/01/zebu-cows-breakthrough-googles-zero.html

*
Hi Bernard (and Roland and Susan), 

Thanks very much for your Stanford Law Codex presentation just now, which I heard from Stanford. Here's the Harvard Law JuryX course with Professor Charles Nesson - https://www.edx.org/course/juryx-deliberations-social-change-harvardx-hls3x-0 - if you're interested; its videos with a history of the jury process are still available for a while. And here's the beginning CC World University and School online law school - http://worlduniversity.wikia.com/wiki/World_University_Law_School - planned in all countries' main languages (since the 10 other beginning law schools part way down), accrediting on CC MIT OCW for Bachelor and Ph.D. degrees (before we move into a new wiki emerging from Wikipedia's database, Wikidata in 358 languages). 

Best, Scott 

*
Hi Daniel, Roland, Susan (Mei and Bernard), 

Thanks very much for your timely and topical Stanford Codex presentation just now, Daniel, which I heard from Stanford. I asked the question about your language strategy thus far re http://congressionaldata.org/ and your https://twitter.com/allcrsreports

Would all of you like to explore developing a CC legal language and translation strategy together and also re World University and School's online law schools, emerging from Wikipedia / Wikidata, and potentially with Google's new Zero Shot translation (https://research.googleblog.com/2016/11/zero-shot-translation-with-googles.html) with WUaS's planned universal translator (for all 7,097 known living languages)? 

Here's the beginning CC World University and School online law school - http://worlduniversity.wikia.com/wiki/World_University_Law_School (before we move into a new wiki emerging from Wikipedia's database, Wikidata in 358 languages) - planned in all countries' main languages (see the 10 other beginning law schools part way down), accrediting on CC MIT OCW in 7 languages and CC Yale OYC for CC Bachelor and Ph.D. degrees.

WUaS would like to build translation and a language/translation strategy (hiring graduate students in many languages to teach the AI in such an Universal Translator) into WUaS in Wikidata as we facilitate translation of CC MIT OCW courses (again now in 7 languages) into all countries' main languages, and for online law schools, too. WUaS would also like to explore collaborating in this with Stanford Law's 5-10 international projects, such as the Stanford Law China Guiding Cases Project (with just the Beijing Supreme Court's ~ 56 cases ). 

How best to bring this project forward - demand progress? - re Stanford Codex, CC WUaS and Congressional Data? 

Thank you.

Best, Scott 

*
In what ways could World University bring together legal translation with Stanford Law Codex re a high bar for generating excellent text versions in other languages, and from HarvardX video too?

*







...



1 comment:

Hammed said...

This is very interesting content! I have thoroughly enjoyed reading your points and have come to the conclusion that you are right about many of them. You are great. Thank you for sharing and keep on update.......

SEO Company in Chennai