English to Portuguese, Spanish, German, Italian, French Translation

Monday, June 26, 2023

History is happening with the free Crow Dictionary App - KULR-TV - Dictionary

[unable to retrieve full-text content]

History is happening with the free Crow Dictionary App KULR-TV

AI is being used to translate 5000 year-old cuneiform tablets - PC Gamer - Translation

With the rise of ChatGPT and the like, AI has entered the mainstream public consciousness. As well as various philosophical concerns or potential economic effects, talk about AI inevitably brings up the topic of the end of civilization.

It may be more relevant to civilization's beginning, however. AI is broadly software that's able to learn, reason, and infer meaning, and that makes it a wonderful tool for translating ancient or dead languages.

A team of archaeologists and computer scientists have created an AI program with exactly that purpose in mind. Specifically, this AI is being used to translate tablets with Akkadian texts using cuneiform script, some of which date back to 2,500 BCE. It's like a super Google Translate.

Akkadian was a language spoken in parts of Mesopotamia, an area now belonging to modern Iraq. According to the authors of the paper published at PNAS Nexus (via Heritage Daily) , there are hundreds of thousands of these clay tablets, but because there's only a limited number of available experts in Akkadian texts, most of them remain untranslated. Mesopotamian languages aren't my strong suit, that's for sure.

The researchers claim the AI is able to achieve 97% accuracy at translating the Akkadian cuneiform script to Latin, which is a much easier task than translating to English, with its more complicated sentence structures.

The AI performed well when tasked with translating formally written text, such as royal decrees or those written by scholars. It does tend to struggle with literary texts, producing what are called "hallucinations", which are results that bear little, if any resemblance to the actual text.

I'd like to see how AIs of the future go at reading some of the rubbish you see on social media today. They'd probably struggle with my PC Gamer articles. Am I rite?

It's hoped that AI will be able to translate other lost languages. With the ability to learn and adapt to the complexities of written text, more knowledge of the ancient world will eventually become available to us.

There's a long way to go though. Even understanding English is a difficult task for AIs. Ordering a Wendy's burger sounds simple enough, but if you've got Cola by Lana Del Ray playing on your radio, you might not get the Sprite you wanted.

Who knows? Maybe one day AI could be used to translate animal speak, like whatever my cat is trying to say. Then again, my cat only really says two things: feed me, or feed me now. I don't need an AI to tell me that.

Adblock test (Why?)

Sunday, June 25, 2023

British Museum apologizes after using translators work in China exhibition without pay or acknowledgment - CNN - Translation

CNN —

When the British Museum launched its “China’s hidden century” exhibition last month, writer and translator Yilin Wang began getting confusing messages from her peers.

The show, which featured 19th century Chinese works including poems by feminist and revolutionary Qiu Jin, didn’t seem to include credits for translators, a friend told Wang. And yet, the Qiu Jin translations seemed to lift directly from Wang’s own work — was she involved in the exhibit?

No, Wang replied: She’d never been contacted by the museum, which used her work without permission, pay or acknowledgment.

A social media firestorm ensued, culminating in the British Museum issuing a statement Thursday that admitted the permissions and acknowledgment for Wang’s translations had been “inadvertently omitted.”

A map of the Qing empire, featured in the British Museum's "China's hidden century" exhibition in London.

James Manning/PA Images/Getty Images

It was an “unintentional human error for which the Museum has apologized to Yilin Wang,” it said, adding that it had removed her translations from the exhibition, and offered payment for the duration they were up, as well as for the translations that remain in a printed catalog.

But these measures fall short and the apology rings hollow, Wang told CNN in a phone interview Friday.

She criticized the statement for sounding passive instead of taking proper accountability. And, she said, it neglects to address the larger questions this incident has raised about ethics in academia and what she describes as the frequent erasure of translators — especially women and people of color.

The firestorm

The online controversy emerged last week when Wang posted about the use of her translations on Twitter.

“Please note this is a copyright infringement … I think you owe me some money for printing and exhibiting my translations, British Museum,” she wrote in a thread, noting that her translations — which had previously been published on her website and in literary journals — were also featured in the museum’s online guide and printed catalog about the exhibit.

Her post has since circulated widely on Twitter, garnering nearly 53,000 likes and 15,000 retweets to date.

The British Museum has since reached out and in its statement Thursday, said it “takes copyright permissions seriously.”

“Across the range of our work, we make every effort to contact the owners of rights in text, images, print and digital media. This was a particularly complicated project and we recognize we made an inadvertent mistake and fell short of our usual standards,” it said.

It added that “China’s hidden century” had involved more than 400 people from 20 countries, and that those involved had “spent years, together with scholars worldwide,” putting it all together.

A British Museum worker pictured in the "China's hidden century" exhibition ahead of its public opening.

James Manning/PA Images/Getty Images

But to Wang, the scope of the project made her erasure sting all the more. “How exactly did this happen?” she said. “It was funded by a research grant that was over 700,000 (British) pounds. These researchers had (almost) four years to research, they must have gathered translations and created all these different formats. It’s been up for multiple weeks, and no one thought to be like, ‘Where are these translations from?’”

The exhibition was supported by a £719,327 ($914,847) research grant from the UK’s Arts and Humanities Research Council called “Cultural Creativity in Qing China 1796-1912.”

#NameTheTranslator

For Wang and peers in the translation and publishing world, this incident highlights the broader and longstanding problem of translators’ work being obscured or uncredited.

A social media campaign known as #NameTheTranslator has picked up steam in recent years, encouraging publishers, educators and reviewers to name translators alongside the original authors of literary works.

“Without translators, these kinds of works would not be accessible,” said Wang, adding that translated works only make up a small minority of books published in the US. “This is especially bad for women translators and women poets.”

The lack of credit also undermines the labor and expertise necessary for effective translation, many translators say. It’s not as simple as running a text through Google Translate — rather, good translation relies on skills, expertise and craft that can take years to train.

“When I’m translating, I am using my knowledge of poetry in English, I’m using my knowledge of classical Chinese literature, I’m doing background research on the poet and … on the time period that Qiu Jin was writing in,” she said. “I am also often going through 10 to 15 drafts of the same poem to find the right words, the right expression, the most eloquent way of translating idioms and allusions, the right way to capture the spirit and emotional power of the poetry rather than a word-by-word translation.”

This can be especially true for classical Chinese, which has a very different syntax and diction from English, she said. So when translations are used without credit, it’s this time, effort and knowledge being poached.

“I would urge the British Museum to come negotiate with me in good faith, that they’d be more apologetic,” Wang said, adding: “It’s really important to have discussions about copyright, about crediting translators’ labor, about making sure that this does not happen again and taking steps to correct it properly.”

The British Museum did not respond immediately to CNN’s request for comment.

Adblock test (Why?)

TWTS: Dictionaries are defined by their editors - Michigan Radio - Dictionary

What’s the name of the book you use to look up words you don’t know?

For many of us, the answer to this question is simply “the dictionary." However, that suggests it doesn’t really matter which dictionary we use to look up a word, and that’s just not true.

Different dictionaries have different approaches, which is why Professor Anne Curzan consults multiple dictionaries when researching your questions.

Curzan was recently at the biennial meeting of the Dictionary Society of North America. The keynote speaker at the conference was David Skinner, author of The Story of Ain’t: America, Its Language, and the Most Controversial Dictionary Ever Published.

The dictionary in question is the 1961 publication of Webster’s Third New International Dictionary. The dictionary had a new editor, Philip Gove, who took a more descriptive approach to the language. Gove took out a lot of usage labels that might have been seen as making judgements about particular words.

When this dictionary was published, it was widely condemned. The New York Times, the Washington Post, and Life Magazine are just a few of the publications that went after this dictionary with phrases like “radically permissive” and “downright irresponsible.” Critics were especially focused on “ain’t” and whether the correct usage labels were used.

The American Heritage Dictionary was actually created in direct response to the controversy over Webster’s Third. Part of that response included a usage panel that Professor Curzan served on. For more on that, listen to the audio above.

Adblock test (Why?)

Saturday, June 24, 2023

5 AI tools for translation - Cointelegraph - Translation

Explore AI translation tools, their features, benefits and pricing models to find the right solution for your translation needs.

Overview

Join us on social networks

Translation is the process of converting written or spoken content from one language to another while preserving its meaning. By automating and enhancing the translation process, artificial intelligence (AI) has significantly contributed to changing the translation industry.

To evaluate and comprehend the structure, syntax and context of the source language and produce correct translations in the target language, AI-powered translation systems use machine learning algorithms and natural language processing techniques.

Types of AI-powered translation systems

AI-powered translation systems can be categorized into two main approaches:

Rule-based machine translation (RBMT)

To translate text, RBMT systems use dictionaries and pre-established linguistic rules. Linguists and other experts create these guidelines and dictionaries that specify how to translate words, phrases and grammatical structures.

While RBMT systems are capable of producing accurate translations for some language pairs, they frequently face limitations due to the complexity and diversity of linguistic systems, which makes them less useful for translations that are more complex.

Statistical machine translation (SMT)

SMT systems employ statistical models that have been developed using sizable bilingual corpora. These algorithms analyze the words and phrases in the source and target languages to find patterns and correlations.

SMT systems are able to make educated assumptions about the ideal translation for a particular input by examining enormous volumes of data. With more training data, SMT systems get more accurate, although they may have trouble with unusual or rare phrases.

Neural machine translation (NMT) has recently become more well-known in the translation industry. To produce translations, NMT systems use deep learning methods, notably neural networks. Compared to earlier methods, these models are better able to represent the context, semantics and complexities of languages. NMT systems have proven to perform better than other technologies, and they are widely employed in many well-known translation services and applications.

Advantages of AI in translation

The use of AI in translation offers several advantages:

Speed and efficiency: AI-powered translation systems can process large volumes of text quickly, accelerating the translation process and improving productivity.
Consistency: AI ensures consistent translations by adhering to predefined rules and learned patterns, reducing errors and discrepancies.
Customization and adaptability: AI models can be fine-tuned and customized for specific domains, terminologies or writing styles, resulting in more accurate and contextually appropriate translations.
Continuous improvement: AI systems can learn from user feedback and update their translation models over time, gradually improving translation quality.

AI tools for translation

There are several AI tools available for translation that leverage machine learning and natural language processing techniques. Here are five popular AI tools for translation:

Google Translate

Google Translate is a widely used AI-powered translation tool. To offer translations for different language pairs, it combines rule-based and neural machine translation models. It offers functionalities for text translation, website translation and even speech-to-text and text-to-speech.

Google Translate offers both free and paid versions. The basic translation services, including text translation, website translation and basic speech-to-text features, are accessible to users for free. However, Google also offers a paid service called Google Translate API for developers and businesses with more extensive translation needs. API usage is subject to pricing based on the number of characters translated.

Microsoft Translator

Another capable AI translation tool is Microsoft Translator. It offers translation services for many different languages and makes use of neural machine translation models. It offers developers APIs and SDKs so they may incorporate translation functionality into their projects.

Microsoft Translator offers a tiered pricing model. It has a free tier that allows users to access basic translation services with certain limitations. Microsoft also provides paid plans for higher volume and advanced features. The pricing is typically based on the number of characters translated or the number of API requests made.

DeepL

DeepL is an AI-driven translation tool known for its high-quality translations. It utilizes neural machine translation models and claims to outperform other popular translation tools in terms of accuracy. DeepL supports multiple language pairs and offers a user-friendly interface.

DeepL offers both free and paid versions. The free version of DeepL allows users to access its translation services with certain usage restrictions. DeepL also offers a subscription-based premium plan called DeepL Pro, which provides additional benefits, such as faster translation speeds, unlimited usage and the ability to integrate the service into other applications.

Systran

Systran is a language technology company that provides AI-powered translation solutions. It offers a range of products and services, including neural machine translation engines, translation APIs and specialized industry solutions. Systran focuses on customization and domain-specific translations.

Pricing for Systran’s offerings is typically based on the specific requirements and level of customization desired by the client.

Trados Enterprise

RWS is a global leader in translation and localization services, and it provides various language technology solutions to support translation and multilingual content management.

One of its language technology offerings is Trados Enterprise (previously RWS Language Cloud). This cloud-based platform is designed to streamline the translation process, enhance collaboration and improve translation quality. It provides a range of features and tools to manage translation projects, such as translation memory, terminology management, project management and linguistic assets.

Trados Enterprise offers different versions tailored to specific needs. The Studio version is priced at $125 per month and provides an industry-leading computer-assisted translation (CAT) tool for professional linguists. The Team version, priced at $185 per user per month, focuses on cloud-based collaboration for translation projects.

The Accelerate version starts at $365 per user per month and offers end-to-end translation management for organizations with custom requirements. RWS also provides a free trial for interested users and encourages potential customers to request a demo to explore their offerings in detail.

Adblock test (Why?)