ShareThis Page

In Pittsburgh, Amazon relies on machines to translate a global market

Aaron Aupperlee
| Tuesday, Feb. 21, 2017, 8:27 p.m.


Translation can be a tricky business.

Just ask John Frazier, Nicole Xu and Lujing Gao, a trio of University of Pittsburgh graduates who last year started uTranslated.

They keep a file of translation miscues, by brands big and small, and it grows daily. In the case of Apple, which learned when it launched its iPhone 7 in China that the Cantonese word for seven also can be used as slang for penis.

And Burger King, whose mango-flavored smoothie in 2014 was translated into PooPoo Smoothie on Chinese-language advertisements.

And whoever was responsible for a sign near the exit of a Chinese subway station that meant to say something like “Please don't crowd the escalator” but instead translated to “Please don't stay.”

“A lot of nuance that's important gets lost in translation,” said Frazier, who rolls out the translation hiccups when he pitches the company.

Translation is a $40 billion industry, Frazier said, and as the world becomes more connected, communicating across languages will become increasingly important.

The trio behind uTranslated want to connect high-quality and fiercely vetted translators with clients needing to take complicated, sensitive or important text from one language to another. Their marketplace for translators cuts out translation agencies — and their fees — and connects clients directly with translators in an effort to cut costs while increasing the pay to translators and the quality of translation, Frazier said.

But then there are companies seeking to cut out humans from the translation equation altogether. Amazon opened an office in South Side Works where most of the 50 employees will work on enabling a computer to translate.

Amazon is using machine translation to provide details and descriptions in its vast online marketplace in the local languages of customers worldwide, Alon Lavie, senior manager of machine translation at Amazon's South Side office, wrote in an email to the Tribune-Review.

Lavie had worked on machine translation at Carnegie Mellon University's Language Technologies Institute and co-founded Safaba Translations Solutions, a university spinoff company, with fellow researcher Robert Olszewski, now a senior technical program manager at Amazon.

“This involves translating millions of words on an ongoing basis, which is infeasible to do with pure human translation,” Lavie wrote, adding that it will take a sophisticated machine translation tool to handle the diversity of Amazon content throughout the company. “Developing high accuracy machine translation systems for this wide diversity of scenarios is a major challenge.”

Lavie said Amazon works in dozens of languages but would not give a specific count of how many. He said a large portion of the Pittsburgh office works on machine translation and other translation technologies.

Alex Rudnicky, a research professor at CMU's Language Technologies Institute, said he believes machine translation technologies are up to Amazon's task. Amazon is trying to get information out in as many languages as possible. The company will not be overly concerned with polished and perfect text, Rudnicky said.

Machine translation creates statistical models that try to predict patterns of words and phrases. Building a machine translation tool starts with taking two identical texts in two languages. Rudnicky said human translated news articles and Wikipedia pages, which volunteers increasingly have translated into several languages, provide troves of data.

The tool analyzes the texts, examining words and the words that often come before and after those words. It builds statistical models around that data, Rudnicky said. It's more than just an automated dictionary.

“You still want a dictionary, but then you need to know how to arrange and rearrange words to make sentences make sense in the different languages,” Rudnicky said. “That's where the statistical models come in.”

The uTranslated founders aren't worried about machines coming for their company. The company, which specializes in English and Chinese translations, has signed major contracts with big brands and recently returned from a trip to China. They met with large publishing groups and other companies. uTranslated took second place in China's Chunhui Cup, a national startup competition, and was recognized with the Award of Excellence in the competition.

“There's always going to be a need for the human touch,” Frazier said. “There are matters of context and subtlety that machine translation will not be able to do. The human touch will never go away from this industry.”

“You can't imagine translating a book through Google Translate,” Xu said. “Especially in Chinese.”

Aaron Aupperlee is a Tribune-Review staff writer. Reach Aupperlee at or 412-336-8448.

TribLIVE commenting policy

You are solely responsible for your comments and by using you agree to our Terms of Service.

We moderate comments. Our goal is to provide substantive commentary for a general readership. By screening submissions, we provide a space where readers can share intelligent and informed commentary that enhances the quality of our news and information.

While most comments will be posted if they are on-topic and not abusive, moderating decisions are subjective. We will make them as carefully and consistently as we can. Because of the volume of reader comments, we cannot review individual moderation decisions with readers.

We value thoughtful comments representing a range of views that make their point quickly and politely. We make an effort to protect discussions from repeated comments either by the same reader or different readers

We follow the same standards for taste as the daily newspaper. A few things we won't tolerate: personal attacks, obscenity, vulgarity, profanity (including expletives and letters followed by dashes), commercial promotion, impersonations, incoherence, proselytizing and SHOUTING. Don't include URLs to Web sites.

We do not edit comments. They are either approved or deleted. We reserve the right to edit a comment that is quoted or excerpted in an article. In this case, we may fix spelling and punctuation.

We welcome strong opinions and criticism of our work, but we don't want comments to become bogged down with discussions of our policies and we will moderate accordingly.

We appreciate it when readers and people quoted in articles or blog posts point out errors of fact or emphasis and will investigate all assertions. But these suggestions should be sent via e-mail. To avoid distracting other readers, we won't publish comments that suggest a correction. Instead, corrections will be made in a blog post or in an article.

click me