Genealogy from the perspective of a member of The Church of Jesus Christ of Latter-day Saints (Mormon, LDS)

Friday, November 18, 2016

How Well Does Google Translate Work for Genealogists?


Because of my language background, I am more than well aware of the problems with translation. This is one reason why I have translation apps on each of my websites. It is also the reason why I am very interested in claimed advances in computerized translation programs. My ideas on instant translation come from Douglas Adams' book, "The Hitchhiker's Guide to the Galaxy."

Adams, Douglas. The Hitchhiker’s Guide to the Galaxy. New York: Harmony Books, 1980.

The Hitchhiker's Guide (which is beginning more and more to look like a smartphone) uses a Babel fish to translate. Here is a description from the radio show.
"The Babel fish is small, yellow, leech-like, and probably the oddest thing in the universe. It feeds on brain wave energy, absorbing all unconscious frequencies and then excreting telepathically a matrix formed from the conscious frequencies and nerve signals picked up from the speech centres of the brain, the practical upshot of which is that if you stick one in your ear, you can instantly understand anything said to you in any form of language: the speech you hear decodes the brain wave matrix."[7]
The Google Translation app for my iPhone translates my spoken speech directly into the target language. I can test this with Spanish since I speak Spanish fluently. It works very well, but not quite as well as the imaginary Babel fish.

In the real world, as I recently noted in a blog post, Google has implemented Neural Networking to its Translate program. Translate has been used by genealogical researchers to aid in working in languages other than those native to the researcher. For a more in depth discussion of the impact of Neural Networking on the Translate program, see the following:

“A Neural Network for Machine Translation, at Production Scale.” Research Blog. Accessed November 18, 2016. https://research.googleblog.com/2016/09/a-neural-network-for-machine.html.
“Google Translate Is Tapping into Neural Networks for Smarter Language Learning.” PCWorld, November 16, 2016. http://www.greenbot.com/article/3142335/android/google-translate-is-tapping-into-neural-networks-for-smarter-language-learning.html.
“Google Translate Just Got a Lot Smarter.” CNET. Accessed November 18, 2016. https://www.cnet.com/news/google-translate-machine-learning-neural-networks/.
“How Google Translate Squeezes Deep Learning onto a Phone.” Research Blog. Accessed November 18, 2016. https://research.googleblog.com/2015/07/how-google-translate-squeezes-deep.html.

 But how well does it really work? I only gave one example in my last post, so I thought I would give a few more examples.

Here is a randomly selected paragraph from The Family History Guide or TheFHGuide.com:
Look for ancestors with a light-blue icon in the couple's box - this means that record sources are available. If you don't see any of these icons, expand one or more ancestor lines to find them. A purple icon means that no sources have been recorded for the person.
Here is the Spanish translation of that same paragraph:
Busque los antepasados con un icono azul claro en la caja de la pareja - esto significa que las fuentes de registro están disponibles. Si no ve ninguno de estos iconos, expanda una o más líneas de ancestro para encontrarlos. Un icono púrpura significa que no se han registrado fuentes para la persona.
This is a pretty good translation. However, one difficulty with Spanish is that there are significant word variations from country to country. I would think, however, that a native speaker of Spanish, who did not know English would probably understand what the Translate program was trying to say.

What about translating back to English from a language I do not know well? Here is a paragraph in Hungarian. I think of a search term I want to use. Then, I translate the term into the language, in this case Hungarian, and use the translated term to search for information. I have Chrome's auto-translate function turned on, so if I want to see the original in the language of origin, I will have to click to see the original.

First the Hungarian:
Anyakönyvek:
- Magyarországi polgári anyakönyvek kutatása itt- Kecskemét, RK anyakönyvek: Ker.: 1678-1920, Ház.: 1737-1922!, Hal.: 1734-1761, stb. itt- Tarnaörs és Erk lakossága kronológiai sorrendben itt- Szlovákia, (Felvidék) egyházi anyakönyvek kutatása itt- Erdélyi települések anyakönyvi adatai itt!- Anyakönyvvezetők magyar-szlovák-német-latin szótára (és más szótárak) itt!- A Magyar Nemzeti Levéltár anyakönyvi mikrofilm mutatója itt- Ausztria, Stájerország : R.K. anyakönyvek a graz-i Püspökség oldalán itt
Then the translation:
Mother Books:
- Hungarian civil registries in research
- Kecskemét, RK registers Ker .: 1678-1920, 1737-1922 .: House Fish !, .: 1734-1761, etc. here
- Tarnaörs and Erk population in chronological order
- Slovakia (Highland) church registers in research
- Birth data Transylvanian villages here!
- Registrar Hungarian-Slovak-German-Latin dictionary (and other dictionaries) here!
- The Hungarian National Archives microfilm ratio at birth
- Austria, Styria R.K. registers of the Diocese Graz side here
You can copy and paste or type in the text you wish to translate. It may not be perfect, but Translate certainly works well enough to get me started in Hungarian research.  

No comments:

Post a Comment