Microsoft’s AI Translates Chinese to English Like Humans
I have used almost every Chinese English translation app out there——some are good, and some are not so good. Translating can get pretty messy fast, and can be quite frustrating when dealing with a lot of text. Google translate is useful when you are trying to translate simple sentences and single words. As long as you don’t try to translate a big chunk of text at once, Google translate does a good job at getting your point across. Over the years, WeChat’s message translation feature has improved greatly, and usually is pretty spot on. If you know your target language well, you can usually hear in your head whether the sentence sounds correct in your translation. But nevertheless, we can all agree to say that translation is not easy.
Well, it looks like Microsoft just might have solved the problem. On Wednesday, a team of Microsoft researchers announced that they’ve created the first machine translation system that’s capable of translating news articles from Chinese to English with the same accuracy as a person. The group says they’ve achieved human parity on a test set of news stories called newstest2017, which was developed by industry and academic partners and made available at a research conference last fall. The test set included about 2,000 sentences from a sample of online newspapers that had been professionally translated.
Many experts had believed designing a computer to match the translation abilities of a human could never be achieved, Microsoft said in a blog post.
“Hitting human parity in a machine translation task is a dream that all of us have had,” said Microsoft technical fellow Xuedong Huang in the post. “We just didn’t realize we’d be able to hit it so soon.”
To create this system, rather than use the conventional translation technique (statistical machine translation), Microsoft used deep neural networks that more closely mimic human behavior. Its team used a dual learning system where the AI would reverse the translation to gauge how well it did. The company also relied on deliberation networks, which mimic the repetitive process people use to revise writing, as well as joint training that used English-to-Chinese translations to boost the training samples for Chinese-to-English work. The AI would even change the direction of its translation (left to right and vice versa) and promote results that were similar both ways.
The researchers caution the system has not yet been tested on real-time news stories, and there are other challenges that still lie ahead before the technology could be commercialized into Microsoft’s products.
If you want to play around with the new translation system, you can do so here: https://translator.microsoft.com/neural. (This is not the production system, and may run slower at times, the site warns.)