This site uses cookies.
Some of these cookies are essential to the operation of the site,
while others help to improve your experience by providing insights into how the site is being used.
For more information, please see the ProZ.com privacy policy.
Post-editing of machine translations is the second-most sought-after skill among language service providers
UOC researchers propose a new method for assessing AI tools to improve the work of translators
Machine translation systems are widely used in the translation industry, although the end results are always reviewed by people (foto: Adobe)
Juan F. Samaniego
ChatGPT and its ability to hold conversations and produce written content have been the focus of a lot of attention in the last year in the field of technology and artificial intelligence. However, AI has been around for some time, helping us in all sorts of everyday tasks, from navigation systems to social network algorithms, not to mention machine translation. Ever since neural machine translation (NMT) systems began to be used on a widespread basis a few years ago, AI has seen exponential growth in its uptake in the translation industry. This has led to new challenges in the relationship between human and machine translators.
Today, the post-editing of machine translations is the second-most sought-after skill among language service providers and is the task with the greatest growth potential, according to the European Language Industry Survey. Translators edit unprocessed machine translations, correcting texts produced by artificial intelligence. This brings with it many advantages for human translators, but also significant problems if the quality of the machine translation is poor. This is why the ability to objectively assess the quality of machine translation tools is essential for the sector.
COLOGNE: Online translation platform DeepL says it launching an AI assistant that, unlike ChatGPT and other rival AI chatbots, is designed to support the writing process with real-time suggestions on word choice, phrasing, style and tone.
The German software company says its new assistant allows users to optimise their texts regardless of their language skills and find the right words for any situation and any reader.
DeepL Write Pro is the company’s first product to be based on its own large AI language model (LLM). LLMs are machine learning models that are trained to understand and generate human language. Well-known LLMs include GPT from OpenAI, Gemini from Google and Llama from the Facebook group Meta.
The new AI writing assistant is primarily aimed at business users and is designed to support teams in companies when writing business content.
DeepL says the assistant, which currently works in English and German, can help companies ensure precise communication from internal content to external customer communications and contracts.
German startup DeepL, known for its AI-powered machine translation, is making a foray into the realm of writing where it adjusts the style of the sentences according to user preference with an emphasis on business application.
The startup launched the service, powered by its own large language model (LLM), in Korea on Friday. It is available in English and German with a plan to expand the number of languages supported.
The style of the sentence can be selected from four style options — business, academic, simple and casual — and four tone options — friendly, diplomatic, confident and enthusiastic.
“Words matter, and language can be the competitive edge that moves the needle for global businesses,” said Jarek Kutylowski, the company’s founder and CEO.
“DeepL Write Pro is our first product powered by our own LLMs, and is the culmination of years of research and innovation that has set us apart from other tech giants.”
In early January 2024, when many in the language industry were likely pondering how to eventually incorporate AI into their offerings or processes, OpenAI announced its GPT Store. Back then, a few translation GPTs could be found, including one built by Phrase called “Phrase Expert.”
By the time OpenAI launched the store to the general public, there were already more than three million GPTs done by pre-release testers. After the kind of hype seen in late 2022 with ChatGPT and all the drama surrounding the company’s CEO and Board in 2023, the store launch was also a popular subject in the news and on social media.
Then, the announcement of the company’s text-to-video generator, Sora, arrived in February 2024, lest people get OpenAI out of their minds for too long.
In March 2024, we asked readers if they had ever used Custom GPTs since the store was launched, and over two-thirds of respondents (64.6%) said No. Over a combined quarter of readers said they tested GPTs a bit after launch (14.6%) or from time to time (12.5%), and a very small percentage said they have been using them daily (8.3%).
Anuvadini, the AI translation tool that is breaking boundaries by translating movie dialogues from Vietnamese to diverse Indian languages will also translate ABK Media’s curated, “The Northeast Connect”, digital content into various Indian languages.
ABK Media, under the mentorship of Dr. Triveni Goswami Mathur, an Educator and Media & Communication Expert, has structured content that delves deep into the heart of Northeast India’s richness, a press release said.
The North East Connect digital content aims to promote cultural understanding and inclusivity. By unravelling the region’s socio-cultural complexities and economic potential, this content will serve as a beacon for scholars, entrepreneurs, and policymakers alike.
Overall, it will equip learners with knowledge to contribute positively to Northeast India’s development.
Anuvadini, will also become the first tool to translate the dialogue of a movie into various Indian languages.
The forthcoming language translation component won’t even require a drawn circle. Google says people will just have to long press the home button or the navigation bar and look for the translate icon. It’ll do the rest. The company showed the tech quickly translating an entire menu with one long press. Google Translate can already do this, though in a slightly different way, but this update means users won’t have to pop out of one app and into another just to check on something.
Google Meet has upgraded its service with AI note-taking and translation capabilities — although it requires an extra subscription cost.
Unveiled at this week’s Google Cloud Next, “Take notes for me” is now in public preview and allows users to engage with the conversation rather than having to manual take down notes or minutes.
Meanwhile, “Translate For Me” is coming in June and will automatically detect and translate captions in Meet. This includes support for 69 languages (equal to 4,600 language pairs), assisting users to feel more confident and connected to their colleagues, regardless of language.
“Workspace customers can already turn on translated captions during video calls and select their preferred language, helping Meet users around the world easily communicate,” said Aparna Pappu, Vice President and General Manager at Google Workspace.
“Statistics and data are what nourishes Machine Translation. But how do you take care of a high-powered machine that, with incorrect information, could cause problems in translation?”
“We have pocket translators and online translation tools in real life, any translation agency would know that they are prone to mistranslation even for simple phrases. But that hasn’t stopped Mark Zuckerberg, the owner of Meta, which manages social media giants like Facebook, and Instagram, to start developing an AI-powered technology that can help us communicate with anyone regardless of their language.”
With many organizations expanding their reach and going global, the role of translation is becoming increasingly important. Businesses need effective multilingual communication with their partners, employees and customers across cultural borders if they are to succeed in international markets.
(…) Some people think that the translating profession’s days are numbered as well. I hear it all the time from non-translators. “Is there even a need for translators when machine translation is easily available?” people sometime ask me when they find out what it is that I do for a living.
Most people have already discovered that machine translation often makes no sense, but they still don’t understand that it will never really make sense because the reason why machine translation often makes no sense is that it is not translation. It is basically just a bunch of words generated by a machine based on an algorithm and it is up to the reader to make sense of these words. More.
Twitter has quietly stopped offering users the ability to instantly translate tweets using Bing’s machine translation feature, slightly more than a year after the company started using Microsoft’s technology.
Users began noticing the feature’s absence earlier this week, though Twitter hasn’t said exactly when it stopped offering the service, or why. But one thing is clear: people who want to get tweets translated from a foreign language will have to copy and paste them into a translation service of choice, rather than clicking a button on Twitter’s website. More.
Up to this point, machine translation (MT) has been improving constantly. As a result of the work, the MT quality is getting better all the time. Ten years ago almost everyone could have said whether the translation was made by a machine or not. Today, between certain languages, simple texts translate so well automatically that it can be nearly impossible to say who made the translation.
To some extent, MT has reached or even exceeded the quality level that can be produced by someone who has studied the language as a hobby. (MT won’t ever be as good as fluent language speakers or translation professionals.) Because the quality is not people have started to use MT more actively. The usage is already so huge that it affects the quality. Whether the effect will be good or bad in the long run is unclear though. Let’s examine the options a little bit closer. More.
Google Translate has hit the 80 languages milestone, as Google announced the addition of support for nine more languages.
“Whether you’re trekking to a new place or simply trying to communicate with someone who doesn’t share a language with you, Google Translate can help you connect to new information and people,” says software engineer Arne Mauser. “Today, we’re launching 9 new languages that span Africa, Asia, and Oceania and have over 200 million native speakers, collectively.”
These languages would be: Hausa (Harshen Hausa), Igbo (Asụsụ Igbo), Yoruba (èdè Yorùbá), Somali (Af-Soomaali), Zulu (isiZulu), Mongolian (Монгол хэл), Nepali (नेपाली), Punjabi language (ਪੰਜਾਬੀ) (Gurmukhi script) and Maori (Te Reo Māori). More.
Contrary to what one might think, Google translate does not work by somehow parsing a text in the source language into its abstract meaning and then restituting that meaning in the target language. Google did not ask language experts to write grammars for every language that would parse them into semantic trees.
Instead, Google uses Statistical Machine Translation (SMT). SMT is based on using large amounts of human translations from language A to language B and uses statistical method to find the sentence it believes is the most likely language B translation of a sentence in language A based on the probabilities derived from that corpus. More.
Mobile apps are international phenomena. So is the Android operating system. Google knows the importance for developers to easily extend the international footprint so today it announced that its App Translation Service is now available to all developers.
The App Translation Service was announced at Google’s I/O developer conference earlier this year and allows for developers to hire human translators across the world to transcribe their apps from English into the local tongue. Want your app to be read in Farsi? Russian? Japanese? Developers can hire a professional translator through Google Play Services to bring their app to users’ native tongue across the world.
Google says that 1.5 million Android devices are activated every day and the vast majority of those smartphones and tablets go to countries where English is not the first language. As such, Google announced at I/O a suite of localization services and techniques to help developers be able to support true international launches of their apps such as the ability to accept a variety of forms of payment and currency through Google Play and to test a localized app. More.
Earlier this week, Microsoft’s research arm showed a project in Asia that have a positive impact on the lives of deaf people – or those who have hearing loss to a greater or lesser degree – who use sign language. The project uses data captured using the Kinect motion control camera usually attached to its Xbox 360 console, and the translates this into words – which can be read or spoken to assist communication with others. It coud even be used to let sign language practitioners issue commands to computers or other devices using sign.
The reverse is also possible, as words spoken by a hearing person can be translated into sign using a digital avatar on a screen (or holographic projection). More
The announcement lists the following uses for the Terminology Service API:
Dictionary-style terminology, definition, and translation lookups for Microsoft terminology on your site or from within your app.
Integration with your localization or content authoring system for dynamic translation and localization.
Combination with Bing Translator to translate your content using lookups to get the right Microsoft terminology and machine translation for the content.
Suggestions for Microsoft terminology in community translation of content and wikis.
Safaba Translation Solutions, LLC, has announced the release of EMTGlobal 3.0, the latest version of the company’s enterprise machine translation (MT) technology. Updates include new language adaptation technologies and a new interface that allows users to align MT service levels with individual business preferences.
Andrea Busse, CEO of IntelaText, and Sue Reager, CEO of @International Services, jointly announced the release of the first speech translation platform for business in 78 languages, distributed under the name Translate Your World. Using this software, a speaker’s voice is automatically translated in real-time. The results appear as subtitles in 78 languages and can be heard as computerized voice in 35 languages.
The Translate Your World suite also includes online video subtitle software, international mobile marketing applications, and crowd-sourced translation management tools. More.
The translation news daily digest is my daily 'signal' to stop work and find out what's going on in the world of translation before heading back into the world at large! It provides a great overview that I could never get on my own.
susan rose (X)
США
ProZ.com Translation News daily digest is an e-mail I always look forward to receiving and enjoy reading!
Share an article »
ProZ.com blog »
