Language Tech Market News
52.7K views | +58 today
Follow
Language Tech Market News
The Home of Multilingual Intelligence
Curated by LT-Innovate
Your new post is loading...
Your new post is loading...
Scooped by LT-Innovate
Scoop.it!

@CondeNast Suggests Fast-TRacking #NeuralMT to Boost @EU Digital Single Market

@CondeNast Suggests Fast-TRacking #NeuralMT to Boost @EU Digital Single Market | Language Tech Market News | Scoop.it
I suggested the EU makes the fast development of neural machine translation one of their flagship technology projects. Typically these EU projects get funding over €1 billion ($1.31 billion) over the course of 10 years. This is what will unleash the digital single market and lead to a similarly explosive growth of content and information as the internet did itself.
more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

Microsoft Translator Enables Engine Customisation for Industry Domains

With Custom Translator, enterprises, app developers, and language service providers can build neural translation systems that understand the terminology used in their own business and industry. The customized translation system will then seamlessly integrate into existing applications, workflows and websites.

more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

SoftBank Ventures Invests $23.5M in Singapore Media Localization Group

SoftBank Ventures Invests $23.5M in Singapore Media Localization Group | Language Tech Market News | Scoop.it

SoftBank Ventures, the early stage venture capital arm of Japan’s SoftBank Group, has invested $23.5 million in Singapore-based localization services provider IYUNO Media Group (dubbing, sub-titling, access services, media processing, and software licensing to studios, network groups, and over-the-top providers). This investment is designed to bring localization technologies to industry applications, including neural MT and voice synthesis foe audio localization and dubbing.

LT-Innovate's insight:

Part of the ongoing shift from boutique suppliers to large-scale tech--driven global video localisation services?

more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

Five Examples of @DeepL Applications

Live translation apps: Recently, tech giants and smaller startups alike have announced the development of earbuds that translate in real time, but can only deliver the quality of traditional translation tools. The new DeepL API allows developers to build live translators with DeepL's translation technology, making far fewer mistakes.
Augmented reality translations: Point your mobile phone camera at a sign and get the right translation. Such an app could be built by any developer with a DeepL Pro subscription.
Operating systems: Translate directly in your browser, chat program, email client, or word processor, as soon as the developers have integrated our technology.
Internal communications: Large corporations can integrate our translation technology into their communication systems. They can then get high-quality translations on the fly with our data protection guarantee ensuring the confidentiality of their information.
Computer-assisted translation: Leading translation tool provider SDL already allows for the integration of the DeepL API into SDL Trados Suite 2017. Every translator can now benefit from DeepL's incredible neural translations within their favorite translation tool.

LT-Innovate's insight:

It will be interesting to see how many language pairs DeepL offers. They appear to be Eurocentric so far. Will this change with the new API service?

more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

@AMPLEXOR Develops Enterprise-Grade #NeuralMT Solution 

@AMPLEXOR Develops Enterprise-Grade #NeuralMT Solution  | Language Tech Market News | Scoop.it

AMPLEXOR is to develop an enterprise-grade, neural machine translation solution targeting the Medical Device market in collaboration with the German Research Center for Artificial Intelligence (DFKI). 

more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

@systran & HarvardNLP To Hold OpenNMT Workshop March 2

@systran & HarvardNLP To Hold OpenNMT Workshop March 2 | Language Tech Market News | Scoop.it

SYSTRAN and HarvardNLP are holding the first OpenNMT Workshop in Paris on March 2nd at Station F, followed by the first ever OpenNMT Hackathon on March 3rd at Télécom ParisTech. Registration is FREE.
The main goals of the workshop are to:
Share different perspectives (developers/users/researchers/industrials) on OpenNMT and Neural Language Processing Applications in general
Present the latest innovations that the community is putting inside the framework
Establish direct contacts within the community
Share results and collaboratively work on a roadmap

LT-Innovate's insight:

This could mean interesting competition between open source NMT groups.

more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

DeepHack.Babel Sign Up Now

DeepHack.Babel Sign Up Now | Language Tech Market News | Scoop.it
А 24/7 week-long sci-hack winter school
LT-Innovate's insight:

MT hackathon to be held in Moscow this month.

more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

Marian - Europe's Latest #NMT Engine

Marian is an efficient Neural Machine Translation framework written in pure C++ with minimal dependencies. It has mainly been developed at the Adam Mickiewicz University in Poznań (AMU) and at the University of Edinburgh. It is currently being deployed in multiple European projects and is the main translation and training engine behind the neural MT launch at the World Intellectual Property Organization.

LT-Innovate's insight:

Available via GitHub.

more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

Lilt Launches Adaptive Neural Machine Translation System

Lilt Launches Adaptive Neural Machine Translation System | Language Tech Market News | Scoop.it

Translation startup, Lilt, has launched an adaptive, neural machine translation (NMT) system using AI and a real-time feedback loop to augment a human translator’s productivity. The system learns from and adapts to its human user enabling more efficient, high-quality translation processes.

LT-Innovate's insight:

Keeping translators in the loop

more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

Unbabel Introduces #NMT-Powered Video Translation Service

Unbabel Introduces #NMT-Powered Video Translation Service | Language Tech Market News | Scoop.it

Unbabel, (US) is launching Unbabel for Video, a new service to provide translations for audio and video content. This  offer a fully managed service powered by Neural Machine Translation, artificial intelligence, quality estimation algorithms, and a global network of more than 50,000 bilingual post-editors. The new service will transcribe and translate video and audio content, returning searchable, time-stamped text in dozens of languages.

LT-Innovate's insight:

Hot area, requiring just-in-time translation using relatively scarce data for the ML algorithms. In other words a data generation service, as well?

more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

New Open Ecosystem for Interchangeable Neural Network Frameworks

New Open Ecosystem for Interchangeable Neural Network Frameworks | Language Tech Market News | Scoop.it
Facebook and Microsoft are today introducing Open Neural Network Exchange (ONNX) format, a standard for representing deep learning models that enables models to be transferred between frameworks. ONNX is the first step toward an open ecosystem where AI developers can easily move between state-of-the-art tools and choose the combination that is best for them.
LT-Innovate's insight:

Should speed up the passage from research to production in machine-learning applications

more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

SYSTRAN Introduces World’s First Automatic Translation System for Classic Korean Literature

SYSTRAN International (Chairman Lucas Ji, www.systrangroup.com) has provided an artificial intelligence-based automatic translation (NMT) solution to Korea's National Information Society Agency(NIA) for translating Korean classics into modern language.

LT-Innovate's insight:

Apparently part of a "neuro-translate the classics" project that could go global. But where's the money?

more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

Neural Machine Translation Research Accelerates Dramatically

Neural Machine Translation Research Accelerates Dramatically | Language Tech Market News | Scoop.it
Five months into the year, the number of papers published in the open-access science archive, arXiv.org, nearly equals the research output for the entire year 2016. The spike confirms a trend Slator reported in late 2016, when we pointed out how NMT steamrolls SMT.
more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

@Unbabel and Microsoft Roll Out New Translation Services

@Unbabel and Microsoft Roll Out New Translation Services | Language Tech Market News | Scoop.it

Unbabel has launched the first “translation-as-a-service” suite for Microsoft Dynamics 365 – bringing AI-powered translation of 28 languages to the company’s set of enterprise resource planning and customer relationship management tools. A beta is available now. The service is powered by a combination of AI and a network of 50,000 human translators. It is the first such offering available on Microsoft AppSource.

more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

New Way to Measure Neural MT Quality?

New Way to Measure Neural MT Quality? | Language Tech Market News | Scoop.it

One Hour Translation (OHT) has recruited, trained, and tested thousands of linguists in over 50 languages, and already run well over 1,000,000 Neural MT rating and testing projects for customers. By the end of April 2018, it will publish the first human-based NMT quality index (initially covering several engines and domains and later expanding), with the goal of promoting the use of NMT across the industry.

LT-Innovate's insight:

Testing MT output may be more time & effort consuming than simply reviewing & post-editing it..

more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

Nvidia Speeds Ups the Hardware Game for #MT with a 16-GPU DGX-2 Server

Nvidia Speeds Ups the Hardware Game for #MT with a 16-GPU DGX-2 Server | Language Tech Market News | Scoop.it

The DGX-2 is a monster. Nvidia has announced that with DGX-2 it has taken the training time of FAIRSeq, a neural machine translation model, from 10 days (on the V100-equipped DGX-1) down to 1.5 days, a 10x improvement in six months. Nvidia claims that it would take an equivalent of about 300 Skylake servers to get that same performance into a single server.

more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

@kantanmt Launches into-Irish #MT Engine for Legal Content

@kantanmt Launches into-Irish #MT Engine for Legal Content | Language Tech Market News | Scoop.it
KantanMT has announced the launch and availability of an Irish Machine Translation engine. Developed by the KantanMT Professional Services team using KantanNeural technology, this engine uses deep learning and neural networks to boost the accuracy and fluency of English to Irish legal translations. The engine has been included as part of KantanFleet, a collection of over 40 Neural Machine Translation engines available to all subscribers of the KantanMT platform.
LT-Innovate's insight:

How long will we need to add the "neural" modifier to the term "machine translation"? All news about MT now is ipso facto about neural.

more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

SYSTRAN Celebrates 50th Birthday With Pure Neural Machine Translation (PNMT) Showcase at AMTA 2018

SYSTRAN Celebrates 50th Birthday With Pure Neural Machine Translation (PNMT) Showcase at AMTA 2018 | Language Tech Market News | Scoop.it
SYSTRAN will be showcasing Pure Neural Machine Technology (PNMT), during the Association for Machine Translation in the Americas 2018 (AMTA) Conference, held March 17 - 21 in Boston, MA.
more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

@SAP Uses #NeuralMT to Solve Multilingual Challenge of Online Learning

@SAP Uses #NeuralMT to Solve Multilingual Challenge of Online Learning | Language Tech Market News | Scoop.it

SAP Globalization Services teamed up with openSAP (MOOC) and used NMT to translate English transcripts into subtitles from the openSAP course, Enterprise Machine Learning in a Nutshell, in December 2017. We received good feedback with over 80% of learners stating that they would like to see more machine-translated subtitles on openSAP. We also saw a large number of learners volunteering to support our future efforts to improve the quality of translations with NMT.

LT-Innovate's insight:

More (expected) kudos for NMT.

more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

#NMT Learns Without a Parallel Data

#NMT Learns Without a Parallel Data | Language Tech Market News | Scoop.it

Two new papers show that neural networks can learn to translate with no parallel texts—a surprising advance that could make documents in many languages more accessible. Imagine that you give one person lots of Chinese books and lots of Arabic books—none of them overlapping—and the person has to learn to translate Chinese to Arabic. That seems impossible, right?” says the first author of one study, Mikel Artetxe, a computer scientist at the University of the Basque Country (UPV) in San Sebastiàn, Spain. “But we show that a computer can do that.”

LT-Innovate's insight:

No translation tech has ever had as many people publishing new investigations on variations of ML in such a short time. Bubble or breakthroughs?

more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

TranslateMedia Boss on Why Many Translators Will Soon Be Editors

TranslateMedia Boss on Why Many Translators Will Soon Be Editors | Language Tech Market News | Scoop.it

(UK) TransMedia's CEO's ’ prediction is that neural MT will change the role of translators. “For the foreseeable future, high-end creative translation will be exclusively human. But, in more and more cases, translators will be asked to act more as editors rather than doing the work from scratch.”

more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

Omniscien Releases Deep Neural #MT Platform

Omniscien Releases Deep Neural #MT Platform | Language Tech Market News | Scoop.it

Dion Wiggins, CTO of Omniscien Technologies: “Only a combination of machine learning technology, effective preprocessing of the training data and data manufacturing to create additional training data can achieve the best results. This is why we have also invested heavily in our entire production workflow, allowing us to build highly customized project specific, high-quality Deep NMT engines in a matter of days. When compared to other offerings in the market, which take around a month to train a standard NMT engine, our solution typically trains an engine within 1-5 days.

more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

Baidu’s Gadget Translates Speech in Near-Real Time

Baidu’s Gadget Translates Speech in Near-Real Time | Language Tech Market News | Scoop.it

Baidu has unveiled a translation device that weighs as much as a smartphone at the 16th Machine Translation Summit in Nagoya, Japan. Using neural MT, it translates speech in near-real time. Currently, it supports Chinese, English, and Japanese, and can also be used as a Wi-Fi hot spot in more than 80 countries — another useful feature for travelers.

more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

Building Your Own Neural Machine Translation System in TensorFlow

Building Your Own Neural Machine Translation System in TensorFlow | Language Tech Market News | Scoop.it

Google has announced a new Neural Machine Translation (NMT) tutorial for TensorFlow that gives readers a full understanding of seq2seq models and shows how to build a competitive translation model from scratch. The tutorial is aimed at making the process as simple as possible, starting with some background knowledge on NMT and walking through code details to build a vanilla system. It then dives into the attention mechanism [3, 4], a key ingredient that allows NMT systems to handle long sentences. Finally, the tutorial provides details on how to replicate key features in the Google’s NMT (GNMT) system [5] to train on multiple GPUs.

LT-Innovate's insight:

Commoditising Neural MT...

more...
No comment yet.
Scooped by LT-Innovate
Scoop.it!

Jonckers Launches Neural #MT Engine

Belgium-based Jonckers' new NMT engine is the latest step in the company’s history of innovation, receiving widespread praise during testing and early launch.

LT-Innovate's insight:

"Widespread praise" is fairly rare for an MT engine!

more...
No comment yet.