Police raid striking doctors' homes, offices, after deadline passes on return-to-work order
Yoon touts improved Japan ties on Independence Movement Day as gateway to 'new world'
S. Korea, US voice 'deep concern' over NK's definition of S. Korea as 'hostile' country
S.Coups of Seventeen gets military exemption due to knee injury
Korean stocks benefit from Zuckerberg's Seoul visit
Vote on bill to probe first lady lays bare Democratic Party split
Address by President Yoon Suk Yeol on the 105th March 1st Independence Movement Day
Cooperation among N. Korea, Russia, China, Iran raises possibility of 'simultaneous conflicts': US general
[Editorial] A country for children
Man jailed for fraud after faking lottery forecast program
Google's AI software wins top score among machines in translation battleBy KH디지털2
Published : Feb. 22, 2017 - 17:03
Artificial intelligence language software by US Internet giant Google Inc., scored higher than its rival AI machines in a translation battle between humans and machines held in South Korea earlier this week, industry officials said Wednesday.
On Tuesday, a group of four professional translators competed against three AI-powered programs provided by Google, South Korea's top Internet provider Naver Inc. and leading automated interpretation company Systran International.
Without revealing names, the organizers said the four professional translators scored better in translating random English articles -- literature and non-literature -- into Korean and other Korean articles into English than the machines.
Of the machines, Google scored a total of 28 out of 60, followed by Naver's automated translation app called Papago with 17 and Systran with 15, the tech company officials with knowledge of the matter said.
"Google's translation level was quite surprising," an industry official who participated in the event said on the condition of anonymity. "It looked like Systran scored the lowest as the company specializes in business translation."
Automatic translation programs using machines have improved recently thanks to the introduction of the technology called the neural machine translation. Major tech giants including Google and Naver rolled out various NMT services last year.
The NMT system is based on a deep learning framework that learns from millions of examples from over 100 different languages.
Unlike previous machine translations that were adopted 10 years ago, the new system considers an entire sentence as one unit.
Previous systems independently translated words and phrases within a sentence. (Yonhap)
Opposition leader retains ticket to his constituency
Car camping: How solo female campers enjoy outdoors
Tensions loom as doctors plan mass rally in deepening clash over med school quota