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.
Every time I tell someone that I’m a Chinese and Italian Foreign Languages major, I watch them do a slight double take. It bothers me, I’ll admit, but not for the reasons you’d expect. I’m frustrated that I can’t seem to conjure a response to their unspoken questions.
“Why learn languages you have no personal connection to? Ones spoken by so few people so far away? Why spend hours memorizing conjugation tables when you’ve already met the language requirements?” In short: “Why learn a new language at all?”
I’ve been studying Chinese for practically my whole life, Italian for a little over a year and French and Nepali during high school and my gap year. And still, I find that the more I throw myself into learning languages, the further I am from conjuring any answer to that simple question: Why?
If you have ever sat through lunch with native speakers at Oldenborg, agonized over verb agreements in office hours or even tried to converse with locals or extended family members in a foreign country, then you already know how it feels to exist in a sort of linguistic exile.
On good days, I feel a mix of pride and exhaustion. On bad days, I feel like setting myself to this impossible task is almost comically akin to Sisyphus’ struggle, except that, unlike him, I have a choice.
I’ve learned that however fluent you may think you are, there is no finish line in learning a language: The mountain remains perpetually stretched out before you, and you must continue to push that boulder further and further up.
Researchers and people from the deaf community have teamed up to co-create a sign language machine translation (SLMT) app.
The research team designed a theatrical performance in sign language, seen through the eyes of artificial intelligence (AI).
“Historically, deaf people have been excluded from the development of automatic translation technologies,” explains Shaun O’Boyle, Research Fellow in the School of Inclusive and Special Education (Dublin City University DCU).
“This has often caused backlash and resistance from deaf communities, as the projects were designed and developed without any input from the very end-users they intended to serve—resulting in a technology no one wanted to use and a big waste of money,” adds Davy Van Landuyt, Project Manager at the European Union of the Deaf (EUD).
For this research, the team decided to reverse the standard approach, with O’Boyle, Van Landuyt, and the other partners of the European project SignON —including the Vlaams GebarentaalCentrum (Flemish Sign Language Centre) — asking participants “If we were to introduce an AI to Shakespeare texts in Irish Sign Language, which extracts would we choose first?
This engagement with the AI allowed them to connect with the audience and gather their opinions about the technology.
In a July 29, 2024 paper, researchers from Apple and the University of Southern California introduced a new approach to addressing gender bias in machine translation (MT) systems.
As the researchers explained, traditional MT systems often default to the most statistically prevalent gender forms in the training data, which can lead to translations that misrepresent the intended meaning and reinforce societal stereotypes. While context sometimes helps determine the appropriate gender, many situations lack sufficient contextual clues, leading to incorrect gender assignments in translations, they added.
To tackle this issue, the researchers developed a method that identifies gender ambiguity in source texts and offers multiple translation alternatives, covering all possible gender combinations (masculine and feminine) for the ambiguous entities.
“Our work advocates and proposes a solution for enabling users to choose from all equally correct translation alternatives,” the researchers said.
For instance, the sentence “The secretary was angry with the boss.” contains two entities — secretary and boss — and could yield four grammatically correct translations in Spanish, depending on the gender assigned to each role.
The researchers emphasized that offering multiple translation alternatives that reflect all valid gender choices is a “reasonable approach.”
Unlike existing methods that operate at the sentence level, this new approach functions at the entity level, allowing for a more nuanced handling of gender-specific references.
The process begins by analyzing the source sentence to identify entities (such as nouns or pronouns) with ambiguous gender references. Once identified, two separate translations are created: one using masculine forms and another one using feminine forms. The final step integrates these translations into a single output that maintains the grammatical integrity of the target language.
AI video creation platform D-ID is the latest company to ship a tool for translating videos into other languages using AI technologies. However, in this case, D-ID also clones the speaker’s voice and changes their lip movements to match the translated words as part of the AI editing process.
The technology stems from D-ID’s earlier work — which you may recall from the viral trend a few years ago where users were animating their older family photos, and later those photos were able to speak. On the back of that success, the startup closed on $25 million in Series B fundraising in 2022 with an eye on serving its increasing number of enterprise customers in the U.S. who were using its technology to make AI-powered videos.
With the company’s now-launched AI Video Translate tech, currently being offered to D-ID subscribers for free, creators can automatically translate their videos into other languages to help them expand their reach. In total, there are 30 languages currently available, including Arabic, Mandarin, Japanese, Hindi, Spanish and French, among others. A D-ID subscription starts at $56 per year for its cheapest plan and the smallest number of credits to use toward AI features and then goes up to $1,293 per year before shifting to enterprise pricing.
D-ID suggests the new AI video technology could help customers save on localization costs when scaling their campaigns to a global audience in areas like marketing, entertainment, and social media. The technology will compete with other solutions for both dubbing and AI video.
In a July 30, 2024 research paper, Otso Haavisto and Robin Welsch from Aalto University presented a web application designed to simplify the process of adapting questionnaires for different languages and cultures.
This tool aims to assist researchers conducting cross-cultural studies, enhancing the quality and efficiency of questionnaire adaptation, while promoting equitable research practices.
Haavisto and Welsch highlighted that translating questionnaires is often costly and “resource-intensive,” requiring multiple independent translators and extensive validation processes. According to the authors, this complexity has led to inequalities in research, particularly in non-English-speaking and low-income regions where access to quality questionnaires is limited.
In questionnaire translation, maintaining semantic similarity is crucial to ensure that the translated version retains the same meaning as the original. As the authors noted, “semantic similarity is more important than word-by-word match.” According to the authors, cultural nuances and colloquial expressions can further complicate this process, making it difficult to achieve accurate translations.
To address these challenges, they developed a web application that allows users to translate questionnaires, edit translations, backtranslate to the source language for comparisons against the original, and receive evaluations of translation quality generated by a large language model (LLM).
In May 2024, researchers emphasized the crucial role that emotions play in human communication and introduced a new dataset designed to enhance speech-to-text and speech-to-speech translation by integrating emotional context into the translation process.
In July 2024, Alibaba incorporated speech emotion recognition (SER) into its FunAudioLLM to retain original emotions in AI-powered interpreting.
Building on this, an August 6, 2024, paper by Charles Brazier and Jean-Luc Rouas from the University of Bordeaux demonstrated how to integrate emotional context into large language models (LLMs) to condition translation and improve quality.
They argue that “conditioning the translation with a specific emotion would use a suitable vocabulary in the translation.”
This research builds on the authors’ previous work, which was the first to explore combining machine translation (MT) models with emotion information. Their earlier study demonstrated that adding emotion-related data to input sentences could enhance translation quality. In this latest study, Brazier and Rouas take the concept further by replacing the MT model used in their prior work with a fine-tuned LLM.
Although event software company Canapii officially launched in 2020, its roots in the events sector stretch back two decades. Initially part of Canalys, a market research company in the IT industry, Canapii’s journey began with the need to automate event processes to minimise human error.
“We started developing an app 18 years ago to streamline our events, initially for internal use,” recalls CEO Rita Chaher. “However, as our sponsors, including tech giants like Lenovo and Microsoft, saw its potential, they requested to use it for their events.”
The pivotal moment came for the Reading, UK-based company in 2019 when Canapii secured a contract to manage several major events. The onset of the pandemic in 2020 necessitated a rapid pivot to virtual events.
“We redeployed our resources, recoding everything to support virtual formats. This shift revealed the tool’s potential beyond the IT industry,” Chaher explains. This adaptation was crucial in establishing Canapii as a standalone entity, focused on both on-site and virtual events.
The relentless pace of software companies rolling out translation AI as a feature continues unabated. Video hosting and sharing services provider, Vimeo, has announced the release of AI-enabled video translation and voice cloning features to its video platform.
The company’s new AI-enabled translation functionality promises to “leverage generative AI to translate video, audio and captions into dozens of languages – while replicating the original speakers’ voices.”
The functionality, which also allows users to add and edit automatically generated captions, is the latest in a series of AI enhancements designed to boost the video-for-business enterprise user base. As part of the release, Vimeo allows enterprise users to generate a free, 30-second sample of the translations, after which it charges on a per-minute basis.
Users can upload videos with multiple speakers and can rate the quality of the translation by selecting a thumbs up or thumbs down icon for each language.
Ashraf Alkarmi, Chief Product Officer at Vimeo said that the company’s AI translation solution “goes beyond simple transcription.” He added that the tool allows users to “maintain the original speaker’s authentic nuances and tone […] in any language.”
The company did not reveal, however, if it is working with a third-party commercial AI translation provider to power the technology.
Trados, the industry-leading translation platform by RWS, has recently launched an innovative ‘AI Essentials’ add-on, simplifying the integration of AI into translation workflows. AI Essentials combines two groundbreaking capabilities introduced earlier this year – Generative Translation and Smart Review – designed to enhance translation quality, speed up time-to-market, and cut translation costs. With no separate LLM subscription required, this add-on empowers organizations to customize their Trados solutions in line with their unique business needs, unlocking new possibilities to translate everything efficiently.
Additionally, we have also released a collection of new enhancements to The Trados platform. Here are just a few of our latest innovations:
Improve quality and reduce turnaround times withbetter translation management
PerfectMatch for efficient translation: We have now introduced PerfectMatch in the browser as a new step in the workflow, bringing this powerful capability to the entire Trados portfolio.
Previously only available in the desktop application, PerfectMatch is a form of context match that compares source files to existing bilingual files rather than to a translation memory. By identifying matching segments from the previous document and considering the surrounding context, it can provide a perfect match.
The Rask AI Platform: A Pioneer in AI Video Translation
In today’s world, as it globalizes, video content creators are under increasing pressure to make their creations available to an ever more diverse international audience. It is now much easier to translate videos with the advent of technology mayflies and Artificial Intelligence (AI) rising. We no longer speak of it as being a difficult or hard task. Innovative tools powered by AI, like the Rask AI Platform, are leading the way in this changed scene.
Video creators can, therefore, extend social media by giving their material actual global reach. Through such platforms, a lot of different features are supported: real-time translation, many video platforms, and localizing videos into multiple languages. This happens most strikingly through these functions, which enable creators to overcome language barriers and reach out to a variety of linguistic and cultural communities around the world.
Best Video Translation Apps
Here is the original list of top AI-driven video translation tools, which are aimed at helping content creators bring their video content to the world. These online video software options are based on the latest new techniques in neural networks and designed right from the metal up to do a full translation that is authentic and fast.
COCONUT CREEK, FL. – MotionPoint, the leading website translation and localization platform, announces the official date of AdaptiveCon 2024! The ultimate event for exploring the latest breakthroughs in translation technology for websites will take place virtually for free on August 7th, 2024, at 12 PM ET.
Why Attend AdaptiveCon 2024:
Cutting-Edge Technology Revealed: MotionPoint will introduce groundbreaking advancements in website translation technology. Don’t miss the opportunity to be among the first to witness these innovations.
Industry Insights: Our event is designed for marketing leaders and web developers eager to stay at the forefront of website translation technology. If you’re looking to harness the latest AI advancements to optimize your translation budget, streamline processes, and enhance quality, this event is for you.
Key Learnings:
Maximize your translation budget by identifying which pages require post-edit translation with groundbreaking AI.
Leverage Brand-Voice AI to improve style-guide adherence, and glossary adherence.
How multinational companies can balance marketing messaging in different countries from Subway’s globalization leader after doing so in over 100 countries.
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.
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.
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.
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.
Share an article »
ProZ.com blog »
