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.
DeepL SE, a well-funded translation software startup that leverages customized artificial intelligence models for improved accuracy over traditional platforms, has announced the debut of its most powerful AI model yet.
The startup’s next-generation language model is said to be designed specifically for translation and editing tasks. It’s based on a highly specialized large language model that’s fine-tuned on enormous amounts of proprietary language data. The company says the LLM provides more “human-like translations” with a reduced risk of hallucinations and misinformation.
The new model’s skills were enhanced via a human model tutoring process that involved thousands of handpicked language experts, who were hired to “tutor” the model to ensure the accuracy of its translations.
Germany-based DeepL has emerged as a rival to better-known translation systems such as Google Translate and general-purpose AI models such as ChatGPT, which can perform translations as part of a much wider repertoire of skills. Because DeepL’s systems are laser-focused on translating and creating business content, the company claims its models provide much more accurate and precise translations for enterprises.
Seán Kelly has called for Irish to be treated equally to other European languages
“Immediate and sustained action” is needed by the European Parliament to ensure Irish is treated equally to other European languages.
That is according to the Irish MEP – and former president of the GAA – Seán Kelly.
He claimed that Irish speakers in Northern Ireland, who hold Irish passports, can face problems in applying for jobs with the parliament.
He has also called for it to establish a link with Queen’s University Belfast (QUB) in order to recruit more Irish-speaking staff.
Mr Kelly has written to the president of the European Parliament, Roberta Metsola, raising his concerns.
Irish language challenges
Irish was recognised as an official language of the European Union (EU) in 2007.
It is one of 24 official languages in the EU, which means that speeches in the parliament and documents are translated to and from Irish.
But since then a shortage of translation staff has meant that the number of EU documents translated into Irish has been limited.
A European Commission report in 2021 said that EU institutions employed 138 Irish speakers, although over half of the jobs were temporary.
It said that although Irish translation had increased, recruiting enough Irish speaking staff was “the main challenge” to providing full services in Irish.
In his English translation of the Hanuman Chalisa, Vikram Seth gives the Chalisa back to those of us who have always loved it and presents it as it should be known and understood to those who are not familiar with it and to those who have been persuaded to misunderstand it.
Representational image: A person dressed as Hanuman for a Ramlila performance. Photo: Satish Krishnamurthy, CC BY 2.0 via Wikimedia Commons
I cannot think of anything better in these times of upheaval and distress than the Hanuman Chalisa given to us in English by Vikram Seth. It was waiting for me when I came home from giving a Ramayana lecture one evening and my heart soared, as if Hanuman himself had come to tell me not to worry, to let him take over the weight of my discomforts and the troubles of the world. He is, after all, the sankata-haran, “crisis dispeller,” as Seth defines him in translation.
The Hanuman Chalisa, Vikram Seth, Speaking Tiger, 2024.
Seth’s translation is pitch perfect – this is no small achievement with a text written in tight couplets with a strict metrical scheme, a text with rhyme and rhythm and a text that is well-known and well-loved. Seth talks extensively in his interviews about the difficulty of confronting the falling, ‘feminine’ metres of the original and finding something equally aurally compelling for the English. You scarcely notice this for Seth’s Chalisa is smooth, bouncy, exuberant, playful, reflecting the sweetness of Hanuman’s nature. But Seth has also found the numen that suggests the presence of divinity and is able to share the overwhelming sense of love and devotion with which the poem is imbued.
The Chalisa is attributed to the Awadhi poet Tulsidas, whose grand composition is the Ramcharitamanas, the story of Lord Ram composed in the 16th century. Tulsidas was a Hanuman bhakt, believing that it was the strong, courageous and big-hearted monkey who was the doorway to Ram’s infinite grace and all-consuming love. For all that Tulsi’s Manas is steeped in devotion to Ram, it is this lyrical and joyous incantation that celebrates the marvellous monkey who is both the paradigmatic bhakt and a metaphor for the human soul’s relationship with the divine. Ram’s perfection might be intimidating, but Hanuman is accessible to everyone – the Chalisa expresses this openness more than any other text.
In a June 19, 2024 paper, researchers from the University of Sheffield, the University of Waterloo, the University of Manchester, the University of International Business and Economics (UIBE), and the tech company 01.AI introduced a multilingual dataset for evaluating machine translation (MT) quality of metaphorical language.
This new dataset aims to fill a gap in MT evaluation by focusing on the complexities of translating metaphors, where the intended meaning differs from the literal interpretation.
Metaphorical expressions pose significant challenges for MT systems because their meaning extends beyond individual words. As the researchers highlighted, “metaphor translation is more challenging than literal translation.”
Despite the fact that “metaphorical expressions are widely used in daily life for communication and vivid description,” the challenge of accurately machine translating them remains largely unaddressed due to resource scarcity and difficulties in handling the variation in linguistic forms and cultural norms inherent in metaphors.
Leader in professional translation services notes that imprecise translation of sensitive documents can put national security at risk by creating misunderstandings of policies and methods.
DALLAS, July 9, 2024 (Newswire.com) – Businesses and government agencies in the defense sector work on a global stage and often need to translate highly sensitive and critical documents, such as operational manuals and confidential communications. ASTA-USA Translation Services, Inc., a leader in professional document translation services, advises military agencies, defense contractors and other organizations that using AI translation services or unvetted translators could put national security at risk.
“Entrusting defense sector translations to AI tools or unverified freelancers poses significant risks. Subject matter expert translators are imperative to ensure precision and cultural sensitivity while safeguarding our national interests and the integrity of our defense operations,” said Alain J. Roy, founder and CEO of ASTA-USA Translation Services, Inc.
Defense industry documents, including manuals, legal documents, and security clearances are highly technical, and every word matters. Generalist translators are insufficient for the task. ASTA-USA always uses translators with industry-specific knowledge, who can speak and write not only in the target language but in the specific terminology of the industry, allowing them to translate specialized defense documents in a source language with precision.
This global convening offers invited artists, writers, and scholars whose work re-imagines the experiences of migration, challenges normative xenophobic ideas and undermines a politics of fear to generate new discourses, aesthetics, and structures of knowledge. Our world is characterized by the movements of people, diasporas, and relocations: far from a temporary or “crisis” phenomenon, human beings globally and historically have always left their homes to escape war, to avoid persecution, for work, for security. We have been uprooted, stolen, trafficked, enslaved; they have been displaced from land despoiled of resources and habitats lost to extreme weather patterns and climate change. We have moved and migrated for deeply private and personal reasons – to reach potential freely, to lead meaningful lives, to secure a future for ourselves and our families. An account of the migration is the totality of many stories. We seek to capture the breadth of experience. Translation is critical to any genuine comprehension of our contemporary world. Literature and art open us to abundant cultures carried by human beings when they migrate. This alternative critical language leads us to think expansively about memory, belonging and identity, familiar and unfamiliar, borders and home, objects and affects, self-imagining, family and loss. Panels include: Perspectives on Migration, Migration and Translation, Black Rest in Translation, MacX Fellows and Migration Projects. We hope you can join us in Paris for what is going to be an unforgettable convening.
Red Cross volunteers getting ready to deliver the mission. Photo illustration by Juan Carlos Molina Padilla/American Red Cross
The American Red Cross Multicultural Communications team announces the launch of its innovative Translator Hub. This cutting-edge platform is designed to bridge the language gap and ensure effective communication with non-English speaking audiences. By utilizing a custom machine translation system complemented by rigorous human reviews, the hub offers accurate and culturally sensitive translations for all text and materials.
The Translator Hub is a significant milestone for the American Red Cross, reflecting its commitment to inclusivity and accessibility. Users within the organization can now translate a wide range of documents, ensuring vital information reaches diverse communities. Whether it’s press releases, social media content, or educational materials, the hub ensures these messages are understood by everyone, regardless of language barriers.
The key to the hub’s success is its dual approach to translation. Initially, a custom machine translation system generates a preliminary document. This is then meticulously reviewed by human translators to guarantee accuracy, contextual relevance and cultural appropriateness. This two-step process ensures that the final output is not only linguistically correct, but also resonates with the target audience.
The need for efficient and effective localization has never been more critical to global success. Organizations operate in an increasingly interconnected world – one that offers immense opportunities but also suffers from an explosion in content.
Traditional localization isn’t built for the scale and pace needed by global enterprises today so struggles to keep up with the ever-growing demand for multilingual content. Each efficiency gain, like translation memory or machine translation, seems to be quickly eclipsed by the sheer volume of content requiring translation. Businesses are often forced to choose between cost, quality and speed. Known as the ‘iron triangle’, it’s long been an insurmountable obstacle to quickly delivering high-quality translations at scale.
To overcome this conundrum, RWS has developed Evolve, its flagship Human+AI solution – building on the philosophy of ‘Genuine Intelligence’ – that aims to redefine what’s possible in the world of localization.
In a June 27, 2024 paper, researchers from Unbabel and Instituto de Telecomunicações introduced xTOWER, a large language model (LLM) designed to generate “high-quality” explanations for translation errors and use them to suggest improved translations.
The researchers explained that machine translation (MT) systems, despite their strong performance, often produce translations with errors. “Understanding these errors can potentially help improve the translation quality and user experience,” they said.
Built on top of TOWERBASE — an LLM designed, trained, and optimized for MT-related tasks —, xTOWER offers detailed, human-readable explanations for translation errors and suggests corrections based on this analysis.
Specifically, the process involves inputting a source text and its translation into xCOMET, which annotates the translation with error spans and assigns a quality score. The complete input (i.e., the source text and its translation), the annotated translation, and the quality score are then passed to xTOWER, which generates explanations for each error span and proposes a new corrected translation based on these explanations.
The Société française des traducteurs (SFT), a French union for professional translators and interpreters, has released a statement on the use of AI translation and GenAI.
The SFT published the statement in mid-June 2024, following a survey of members between November-December 2023, and now encourages all professionals — members and the unaffiliated alike — to share the statement. (The SFT reported 1,648 members as of June 2020.)
The statement acknowledges that the translation industry has already been impacted by major technological upheavals; namely, the introduction of neural machine translation (MT) in 2016. Since then, clients and language service providers (LSPs) alike have largely adopted this technology.
Still, the statement posits, that the output [of machine translation] remains unreadable in its raw state, and requires humans to correct it via post-editing: “But 70% of our member translators who responded to our survey considered PE (and by extension AI) a threat to their profession.”
More specifically, the SFT calls for the respect of human expertise and strongly recommends against replacing human language experts with AI tools, particularly in high-stakes scenarios.
In a June 12, 2024 paper researchers from Tencent AI and the Harbin Institute of Technology introduced TasTe, a method for teaching large language models (LLMs) to translate through self-reflection.
The key idea is to enable LLMs to generate preliminary translations (i.e., drafts), self-evaluate their own translations, and make refinements based on the evaluation.
The researchers explained that LLMs have shown exceptional performance across various natural language processing tasks, including machine translation (MT). However, their translations still do not match the quality of supervised neural machine translation (NMT) systems.
To address this, the authors proposed the TasTe framework (translating through self-reflection), which improves the translation capabilities of LLMs by incorporating a self-reflection process.
Minors who were locked up in adult prisons for people smuggling say they could not understand proceedings and thought they were going home.
Vulnerable Indonesian children say they were either given no interpreter or an interpreter who spoke the wrong language during deeply flawed people smuggling prosecutions, leaving them unable to understand court proceedings before their imprisonment by Australia in maximum security adult jails.
The Australian government last year agreed to pay $27.5m in compensation to more than 200 Indonesians who were wrongfully prosecuted and detained as adult people smugglers while they were children.
The children were found on boats carrying asylum seekers to Australia in 2009 and the early 2010s during the last Labor government.
The majority were from poor fishing villages and many were tricked into joining the voyages by people-smuggling ringleaders.
Durrow man translates Heartstopper – the queer love story everyone is falling for. – into Irish.
Tadhg Mac Dhonnagáin (foilsitheoir) James Kavanagh (a sheol an leabhar), Eoin Mc Evoy (aistritheoir)
It was a full house in An Siopa Leabhar, 6 Harcourt Street, Dublin 2 on Thursday evening for the launch of Heartstopper – the queer love story everyone is falling for.
Written by Alice Oseman, the series is a global hit on Netflix and the books are available in more than 36 languages.
And now, thanks to the superb translation by Eoin McEvoy, the first book in the best-selling series of graphic novels is also available as Gaeilge.
From Durrow, Eoin is a translations and creative writing lecturer in the School of Irish, Celtic Studies and Folklore, UCD.
Google LLC said today that it’s bringing 110 new languages to its web and smartphone translation app using the power of artificial intelligence, making it more comprehensive than ever with 243 languages in total.
This is the largest expansion to date for Google Translate since 2022, when the company brought 24 new languages to the app using zero-shot machine translation. That’s where a language model learns to translate a language without ever seeing an example.
The company employed PaLM 2, a transformer-based large language model AI developed by Google Research that first powered Bard, Google’s AI chatbot. It eventually evolved into Gemini, which is now powered by the company’s AI model of the same name. Differing from Gemini, PaLM 2 was trained on Pathways, a vast dataset of human language containing more than 1.56 trillion words and 250 billion parameters.
Given the size of this dataset, Google said, PaLM 2 can attain unprecedented fluency with written languages and demonstrated an impressive ability to perform linguistic tasks during testing including understanding idiomatic phrases. However, unlike Gemini, PaLM 2 cannot understand or generate images or work with audio.
Copenhagen, Denmark – June 27 2024 – EasyTranslate, an AI-powered language operations platform, has launched HumanAI, designed to transform the translation industry.
The translation market has long been split into two segments: technology providers and service providers. This separation has created a gap, as there has not been a single solution that combines both. The rise of the LangOps Platform addresses this issue by integrating services and technology into one cohesive solution, eliminating the need for separate TMS and service providers. This innovation fills a crucial gap in the market, streamlining multilingual content management and service delivery in one cohesive solution – Language Operations Platform.
HumanAI combines advanced AI models with strategic human intervention to deliver unparalleled quality, speed and cost-effectiveness, making it a powerful tool for businesses looking to improve on-brand global communication.
In the latest edition of Google’s SEO Office Hours video series, a company representative gave an update on the search giant’s take on AI-generated translations for multilingual websites.
The series, led by Google’s Search Relations Team Lead, John Mueller, answers questions from users about website crawling, indexing, and internationalization.
One user asked, “how can one be transparent in the use of AI translations without being punished for AI-heavy content?” Google’s Mueller responded that while there is “no special markup” to label webpages as automatic translations, users should “consider whether translated pages align with the quality bar that you set for yourself.”
Mueller suggested that if the translated content isn’t high quality, users would be better off not indexing the pages for search engines. “Ultimately, a good localization is much more than just a translation of words and sentences, so I would definitely encourage you to go beyond the minimal bar if you want users in other regions to cherish your site”, he concluded.
It will soon be easier to see Facebook and Instagram posts in lesser-spoken global languages, but an expert suggests that to improve the tool Meta should talk to native speakers.
It will soon be easier to see Facebook and Instagram posts in 200 lesser-spoken languages around the world.
Meta’s No Language Left Behind (NLLB) project announced in a paper published this month that they’ve scaled their original technology.
The project includes a dozen “low resource” European languages, like Scottish Gaelic, Galician, Irish, Lingurian, Bosnian, Icelandic and Welsh.
According to Meta, that’s a language that has less than one million sentences in data that can be used.
Experts say that to improve the service, Meta should consult with native speakers and language specialists as the tool still needs work.
Evriviadis Sofos received the best translation award. Credit: LEA
Translator Evriviadis Sofos received the best translation award at the 16th LEA (Literature in Athens) Festival which aims to form cultural and literary bridges between Latin America, Spain, Greece and Portugal through words and letters.
Organized under the auspices of the Greek President Katerina Sakellaropoulou, and supported by the Libra Group and Libra Philanthropies, the event was attended by guests from Argentina, Chile, Colombia, Spain, Mexico, Nicaragua, Peru, Portugal, Puerto Rico and Uruguay.
At the main festival event which place earlier this week at the Amphitheater of the Acropolis Museum, Sofos was awarded for his translation from Catalan of the book “Uncertain Glory” by Spanish writer Joan Sales, published by Agra Editions in 2023.
The LEA Festival inaugurated the literary translation prize in 2021 with a symbolic prize of 1,000 euros. This year, 35 books were submitted to the LEA festival for judging.
In a June 4, 2924 paper researchers from Google Research and Google DeepMind presented SimulTron, a model designed for on-device, real-time speech-to-speech translation (S2ST) built upon the Translatotron architecture.
The researchers highlighted the ongoing evolution of S2ST technology while emphasizing the persistent challenge of achieving accurate, real-time, on-device simultaneous translation. They noted that existing simultaneous translation models are not adequately optimized for the unique constraints of mobile devices, underscoring the significant challenge of achieving seamless real-time translation on mobile devices.
“Today, with smartphones and tablets being central hubs for personal and professional interactions, on-device S2ST is crucial,” they said.
In response to this challenge, they introduced SimulTron, a model that leverages the strengths of Translatotron while incorporating key modifications specifically tailored for the on-device, simultaneous translation scenarios.
Earlier this year, a drumbeat of news headlines played into public anxieties about the safety of human jobs when Duolingo, a language learning app, became a prominent example of a company cutting workers and replacing them with artificial intelligence.
The most eye-catching job cuts were those for translators, who worked on some of the company’s less popular language education courses. Translators and interpreters are often near the top of media listicles as the jobs most likely to be killed by AI. When the stories about Duolingo’s job cuts circulated, they seemed to confirm that the inevitable AI jobs apocalypse had arrived.
In a recent conversation with Planet Money, the CEO of Duolingo, Luis von Ahn, downplayed the meaning of the cuts. It wasn’t full-time employees. It was only 10% of their contractors. His company’s recent embrace of generative AI only played one part in the decision, and so on. More interesting, considering Duolingo’s official partnership with OpenAI, was von Ahn’s reaction to the company’s recent demonstration of its newest version of ChatGPT, GPT-4o.
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)
ארצות הברית
I receive the daily digest, some interesting and many times useful articles!
Share an article »
ProZ.com blog »
