A Real World Example of Machine Translation
by Daire Coco
A few weeks ago I received my first official communication using machine translation. As a translator, I use MT all the time, but this was the first time I had ever been the end user of it. At the risk of sounding like a total geek, I have to say it was an exciting moment.
The machine translation was included in an email I received from Kiva, a microloan organization I joined about five years ago. The email was an update about a woman in Honduras named Elena, who had used the money that I, and other Kiva members, had loaned her to start her own business:
- Elena is very happy and grateful that thanks to the financial support provided by the investors in the Kiva platform in partnership with the cooperative COMIXMUL she could apply agricultural inputs, skilled pay for cleaning work and achieve get better production . With this financial support she improved their living conditions better income generating at home for the sustainability of the whole family. What I like most about working with COMIXMUL is the support provided by the credit and the benefits that it provides, the treatment staff is very nice and other services that help overcome. Your goal is to keep fighting to get ahead for a better life and a better economic position and give your family a better stability in terms of each of their needs.
The MT gets off to a good start, but quickly deteriorates. Agricultural inputs? Skilled pay? Achieve get better production? Services that help overcome? The inexplicable shift from the third person to the first and second. Well, at least we know Elena is happy and grateful for the financial support provided by Kiva investors, which is the main point of the text.
In addition to the unedited MT in English, Kiva provided a link in the email to the original Spanish version of the text. If you speak Spanish, you’ll see that the mistakes in the computer translated text could have been fixed with minimal post editing:
- Doña Elena se encuentra muy contenta y agradecida porque gracias al apoyo financiero brindado por los inversionistas en la plataforma de Kiva en alianza con la cooperativa COMIXMUL ella pudo aplicar los insumos agrícolas, pagar mano de obra calificada para la limpieza y así lograr obtener una mejor producción. Con este apoyo financiero ella mejoro su condición de vida generando mejores ingresos en su hogar para la sostenibilidad de toda su familia. Lo que más le gusta de trabajar con COMIXMUL es el apoyo brindado por medio del crédito y los beneficios que la misma brinda, el trato del personal que es muy agradable entre otros servicios que ayudan a superarse. Su meta es seguir luchando para salir adelante para tener una mejor vida y una mejor posición económica y brindar a su familia una mejor estabilidad en cuanto a cada una de sus necesidades.
At the end of the email, Kiva asked for feedback about the update:
As I mentioned, the machine translation was exciting to receive in a geeky translator sort of way. But was it exciting to read? No, it was annoying to read in a I-have-no-idea-what-this-means sort of way. When I clicked on Mr. Frowny Face, a box popped up asking me to comment, so I wrote something to the effect of: “Machine translation is a great tool when used in conjunction with a human translator.” In other words, this is not a Google driverless car. You still need humans in MT, even if they are backseat drivers.
Although I may sound critical, I applaud Kiva’s efforts to employ technology as a way to help connect its members with the people they support. It’s a step in the right direction. Kiva just needs to follow through with the next step and start using its team of volunteer translators as post editors of machine translation. This is a great example of where MT works well and an even better illustration of why I strongly encourage other translators to learn how to edit and post-edit, if they don’t already.
If you’d like to help out Kiva while sharpening your translating or editing skills, click here.