Flitto’s language data helps machine translation systems get more accurate
Artificial intelligence translation tools are constantly improving but are not yet perfect. Data is required to train these tools, which machines themselves cannot create*. This is where Flitto comes in, offering what they call corpus.
“Flitto’s corpus includes sets of human-translated sentences from its crowdsourcing service, which is used for things like slang, pop culture references or dialects that might stymie a machine translation service. Over the last five years, Lee says Flitto has accumulated more than 100 million sets of translated language data.”
Amongst their customers are Baidu, Microsoft and NTT DoCoMo.
*some recent developments show otherwise, see article below
Source: TECHCRUNCH
Researchers Have Created an AI System That Teaches Itself New Languages
Researchers have demonstrated that AI systems no longer require human translated content to learn from:
“They hinge upon the way that words are connected in similar ways across different languages – for instance, ‘table’ and ‘chair’ are frequently used together, no matter the dialect. By mapping out these connections for each language and then comparing them, it’s possible to get a decent idea of which terms relate to one another. This process is not supervised by a human.”
There is a still room for improvement, and introducing semi-supervised systems could ease the process.
Source: FUTURISM
AWS launches translation services
Amazon has just announced the launch of their new translation service based on language pairing models to benefit businesses’ international expansion.
“The model consists of an encoder component which reads sentences from the source language and creates a representation that captures the meaning of the text provided. The model also has a decoder component that formulates a semantic representation used to generate a translation of the text from the source language to the target language. In addition, attention mechanisms are used by the service to build context from each word of the source text provided in order to decide which words are appropriate for generating the next target word.”
Source: TECHCRUNCH
Google Translate’s gender bias pairs “he” with “hardworking” and “she” with lazy, and other examples
Certain languages employ gender neutral pronouns that when translated into English, Google Translate needs to guess the gender, highlighting biases in our language.
“The algorithm is basing its translations on a huge corpus of human language, so it is merely reflecting a bias that already exists. […] At the same time, automation can reinforce biases, by making them readily available and giving them an air of mathematical precision.”
Source: QUARTZ
Microsoft built an AI-powered iOS app to help you learn Chinese
The main purpose of this AI-powered iOS app is for users to get a chance to practice Chinese when human communicators are not available.
“As you use the app, you’ll get scored on your speaking ability and highlighted words that you need to work on, plus sample audio to hear how the words are actually pronounced.”
Source: ENGADGET
Google Lens receives its first update w/ image translation and saving, Google Keep import
In addition to adding the “Import to Keep” and “Remember this” features, Google Lens has integrated the Image Translation feature.
“Tapping or verbally saying “Translate this” begins the process, with translated words overlaid over the image rather, thus preserving context as best as possible.”
Source: 9TO5GOOGLE
This article is part of our Weekly Roundup, where we provide you with the latest news on Localization, Translation, and Language. For questions or comments, please click here.