[unable to retrieve full-text content]
Dragon Ball Super Sparks Debate Over Chapter 101 Translation Blip ComicBook.comSunday, January 21, 2024
Next-gen AI-powered voice translator previewed by Vasco at CES - Fox Business - Translation
Making Money panelists Shah Gilani and Scott Martin reveal how Apple can craft excitement in the marketplace and break down the price of oil.
Vasco Electronics previewed its next-generation live voice translation device that is powered by artificial intelligence (AI) and can provide live translations of nearly 50 languages to your ear.
The company, headquartered in Poland, previewed the device at CES 2024 in Las Vegas last week. The Vasco Translator E1 uses earpieces that are connected to a phone app and can translate 49 languages in real-time with an audio translation that someone can hear through the earpiece. The translation also appears as text in the app for convenience.
The Translator E1 can accommodate conversations with up to 10 people when a mobile app is used. The tool allows each user to speak their own language and hear the response in that language. The earbuds fit over the ear, rather than in the ear, for hygienic purposes, and it can be used with two earpieces or one earpiece and a phone.
Tomasz Stomski, Vasco’s chief product officer, told FOX Business at CES that the E1 was designed to be more user-friendly for longer conversations than its Translator V4, which is more useful for a "fast conversation, like if you’re traveling or need to get something done quickly."
WHAT IS ARTIFICIAL INTELLIGENCE (AI)?
The Vasco E1 Translator earbud. (Courtesy of Vasco Electronics)
"We decided to go for something that is more natural so you don’t have to hold the device in your hand, you can put it on the table and put on your earbuds and set it to touchless translation mode and have a conversation," Stomski said.
Vasco currently sells the Translator V4, which is a handheld device that resembles a smartphone and provides live translation of conversations and can also be used to translate text from images taken by the user. The V4 can provide speech translation in 76 languages, in addition to photo translations in 108 languages and text translation in 90 languages – though it is most useful for translating shorter conversations.
Stomski said Vasco wants the translation tools to meet the needs of customers who travel regularly, as well as those who are ex-pats, working in international teams or are in families that have language barriers due to relatives being from different countries.
SAMSUNG UNVEILING ‘NEW ERA OF MOBILE AI,’ S24 LINEUP DURING GALAXY UNPACKED EVENT
The Vasco E1 Translator with earbuds in the case. (Courtesy of Vasco Electronics / Fox News)
The Vasco Translator E1 is expected to be available in the U.S. in the second quarter of this year, while it will be available this March in Europe. It’s also compatible with the Translator V4 for users who wish to use both devices.
Stomski explained that AI is used to help identify human voices to filter out background noise that could otherwise interfere with the live translation or the transcript generated by the devices.
"All of the translations there are connections with AI of course, but starting with the earbud – here we’ve got a model that helps us to recognize the human voice because we’re checking what is going into the microphone and we’re checking with the AI model if this is a human voice or this is like a car or a dog," he said.
The Vasco V4 Translator. (Courtesy of Vasco Electronics / Fox News)
GET FOX BUSINESS ON THE GO BY CLICKING HERE
Stomski added that the device uses several models from different providers, in addition to in-house solutions to cross-check speech translations as well as the image-to-text translations that Vasco’s tools can provide.
"We are constantly checking the quality of the translation for every pair because we cannot do it in a common language," he explained. Those translations are sent to Vasco’s cloud, which helps semi-automatically test the roughly 6,000 pairs of languages covered by the devices before those results are checked by humans.
Inside DeepL: How the startup is rivalling machine translation giants - TNW - Translation
AI translation has transformed the way we communicate, breaking down language barriers in an unprecedented way. The sector’s global market size is projected to reach $12.3bn (€11.3bn) by 2026 — and big and small players alike are aiming to cash in.
Among them, Cologne-based DeepL has been raising industry standards even compared to tech giants like Google Translate and Microsoft Translator.
“We’ve always been rivalling big companies.
The startup was born from online dictionary Linguee and has grown fast since its founding in 2017 by Jarek Kutylowski, a computer scientist who’s also serving as the company’s CEO.
TNW Conference 2024 - Speakers announced!
Meet the powerhouse experts that will take the stage on June 20 & 21 in Amsterdam and save your seat today.
Born in Poland, Kutylowski moved to Germany at the age of 12 where he attended school without speaking a word of German. This made him realise the importance of language and the difficulty in having to communicate in more than one.
When he began working on DeepL in 2017, he saw that neural networks might offer a breakthrough that would enable technology to solve this problem. “We kind of knew that machine translation was going to go in this direction. And we knew this was going to be immensely helpful. Seeing this opportunity, we thought ‘Hey, let’s build something great,’” Kutylowski tells TNW.
Neural Machine Translation (NMT) — as in, the one using neural networks — is the most successful machine translation method we have to date. Compared to its predecessors, it’s faster, more accurate, less resource-intensive, and easier to scale.
DeepL uses the technology to offer free and premium translation services, with special focus on B2B products. It says that, since its inception, over 1 billion users have made requests, and that it currently has more than 20,000 business customers, including the likes of Elsevier, Fujitsu, and Mastodon.
“Translation is really important for businesses,” Kutylowski explains. “Nowadays, companies start going global and expanding into other markets very quickly, so they get customers in different areas.”
He adds that the largest need for translation lies usually in those professions that are text-heavy, such as legal services. “This is where we see most often the strongest demand from our customer base,” Kutylowski says.
To date, DeepL covers 31 languages spanning across Europe and Asia. In 2023, it introduced its AI writing companion and secured unicorn status. Despite the tough funding environment, in January the company raised an undisclosed amount (estimated at €100mn), reaching a €1bn valuation.
The “world’s best” machine translation
DeepL makes the confident claim of offering the “world’s best” AI translation. In addition, it says its product is more nuanced and 3x more accurate compared to those of its competitors.
These assertions are based on “blind tests,” in which professional translators select the most accurate translation without knowing which company produced it.
When I ran a few test experiments of my own, DeepL did indeed come out on top. Firstly, I used a passage from The Stranger by Albert Camus and translated it from French into English, using bothDeepL’s translator and Google Translate.
Although the literary text category isn’t the purpose that these tools have been built for, I decided to begin with it anyway, as it is by default more difficult for an AI system.
That’s because the art of literary translation is complex and requires more than just linguistic proficiency. It involves a high level of creativity, a deep understanding of the author’s voice, style, and socio-historical background, as well as the transfer of meaning across different cultures.
Nevertheless, DeepL’s result was by far superior to Google’s. While it missed some uses of metaphorical language and made a few errors of intent and agreement, the end text did provide a closer meaning to the original.
I repeated the experiment using an article of my own to examine whether the translation tools conveyed the meaning I myself had intended. I translated from English into Greek (my native language).
Below is the DeepL translation:
And here is the Google Translate result:
Again, DeepL did a better job. Despite a couple of minor mistakes, the translation was more nuanced and natural in Greek, while also sticking to the original meaning. But since that’s probably all Greek to you, you don’t have to take my word for it, test it for yourself.
According to Kutylowski, conveying the right meaning to the target language without “butchering” it requires the right balance between accuracy and fluency. This heavily depends on the use case. For example, a technical document calls for higher accuracy, while a marketing text for higher fluency.
Despite this challenge, he believes that AI is capable of learning even the most complex languages. “If there was an alien language that we had to learn, nowadays, with the proper amount of material translated, we could probably work out the translation model for that too,” he adds.
What’s DeepL’s edge?
Kutylowski doesn’t seem concerned about the competition. “We’ve always been rivalling big companies,” he notes, adding that Google Translate remains DeepL’s biggest competitor.
He says that the startup’s edge comes down to a combination of three factors: hard work, a great team, and focus.
“Focus is always an important thing,” he says. “Translate isn’t the core business of Google — it’s one of the 100 side gigs. The same goes if you consider LLMs and the OpenAIs of this world as our competition; translation is only one thing of what they’re doing and their GPU is doing a tonne of different things. We’re focused on one particular area.”
From a technological perspective, DeepL’s success lies in the architecture of its neural networks, the input from human editors, and the training data.
The startup trains its models on tons of data, mostly from the internet, and employs special web crawlers to automatically find translations and assess their quality. It also uses methods such as reinforcement learning to provide positive feedback to the AI so it keeps producing the desired quality.
It’s also about finding the right balance between the capability of the model to translate and its capability to form sentences in the target language, Kutylowski adds. “So a lot of work goes into how much we are training the models on monolingual data and how much on translations. There are a lot of details which the mathematical team is taking care of.”
Machine translation: new challenges and opportunities
Kutylowski acknowledges that the recent boom in fascination with AI — to a great extent because of Large Language Models (LLMs) — has resulted in a more challenging and fast-paced landscape.
“Machine translation is a race now.
DeepL’s team now has to keep up to date with multiple developments: new models coming out, the open-source work that’s happening, academic research, and the work of other companies.
“Machine translation is kind of a race right now,” he says. So what’s a good strategy for competing in this race?
According to Kutylowski, one aspect is to be continuously innovating, and ensuring you’re taking the right steps to enable that. Investing appropriately is another. It also comes to securing the required capital and the right team.
But at the same time, the exponentially increasing interest and advancement in AI also brings new technological opportunities. “There are things that we thought about two or three years ago that technology wasn’t there yet to enable,” he says.
This includes personalised translations that fit a company’s style and a more interactive translation experience. DeepL has also begun research into invoice translation, while it’s training its own LLMs from scratch — in part, thanks to its new supercomputer cluster, DeepL Mercury.
These LLMs will open up opportunities to further improve translation quality and enable new, interactive workflows for users, with more capabilities and applications to be unveiled in 2024.
The future of language learning
Machine translation has had a tremendous impact on overcoming communication barriers. But this also bids the question: will we reach a point where we’ll no longer be learning foreign languages because AI can do it for us?
“With AI advancement in general, I think we as humans will have to ask ourselves the question, what do we need to learn? And what do we want to learn?,” Kutylowski says.
He believes that, when it comes to surviving in a foreign country, the necessity of language learning will gradually decrease as the technology gets better and better. But this doesn’t mean that the value in learning languages will decrease.
To illustrate, he uses maths as an example. While in real life we don’t apply the majority of the complex equations we were taught at school, the process of learning is still vital because it teaches us rational thinking.
The same goes for languages, Kutylowski says. When we learn a language we are also learning how to form thoughts and articulate ideas — and that’s crucial for our development.
The benefits of foreign language learning and bilingualism are indeed far-reaching both for the individual and the society.
Research shows that learning a second language actually changes the brain. Specifically, it increases the density of grey matter (the number of neurons in the brain) as well as the integrity of white matter (the system of nerve fibres that connect the different regions of the brain). This not only strengthens overall brain function, but also enhances memory, attention, concentration, and other cognitive abilities.
In addition, numerous studies have linked language learning to better academic performance, higher employability, improved creativity, as well as communication skills and cross-cultural awareness.
“So for your own pleasure and for the development of your brain and personality, it’s still going to be important to learn languages,” Kutylowski notes. “And even with the best translator on your phone, if you’re marrying a partner who’s from a different country, you’re not going to be communicating through your phone. Or at least, I hope so.”
Saturday, January 20, 2024
Next-gen AI-powered voice translator previewed by Vasco at CES - Fox Business - Translation
Making Money panelists Shah Gilani and Scott Martin reveal how Apple can craft excitement in the marketplace and break down the price of oil.
Vasco Electronics previewed its next-generation live voice translation device that is powered by artificial intelligence (AI) and can provide live translations of nearly 50 languages to your ear.
The company, headquartered in Poland, previewed the device at CES 2024 in Las Vegas last week. The Vasco Translator E1 uses earpieces that are connected to a phone app and can translate 49 languages in real-time with an audio translation that someone can hear through the earpiece. The translation also appears as text in the app for convenience.
The Translator E1 can accommodate conversations with up to 10 people when a mobile app is used. The tool allows each user to speak their own language and hear the response in that language. The earbuds fit over the ear, rather than in the ear, for hygienic purposes, and it can be used with two earpieces or one earpiece and a phone.
Tomasz Stomski, Vasco’s chief product officer, told FOX Business at CES that the E1 was designed to be more user-friendly for longer conversations than its Translator V4, which is more useful for a "fast conversation, like if you’re traveling or need to get something done quickly."
WHAT IS ARTIFICIAL INTELLIGENCE (AI)?
The Vasco E1 Translator earbud. (Courtesy of Vasco Electronics)
"We decided to go for something that is more natural so you don’t have to hold the device in your hand, you can put it on the table and put on your earbuds and set it to touchless translation mode and have a conversation," Stomski said.
Vasco currently sells the Translator V4, which is a handheld device that resembles a smartphone and provides live translation of conversations and can also be used to translate text from images taken by the user. The V4 can provide speech translation in 76 languages, in addition to photo translations in 108 languages and text translation in 90 languages – though it is most useful for translating shorter conversations.
Stomski said Vasco wants the translation tools to meet the needs of customers who travel regularly, as well as those who are ex-pats, working in international teams or are in families that have language barriers due to relatives being from different countries.
SAMSUNG UNVEILING ‘NEW ERA OF MOBILE AI,’ S24 LINEUP DURING GALAXY UNPACKED EVENT
The Vasco E1 Translator with earbuds in the case. (Courtesy of Vasco Electronics / Fox News)
The Vasco Translator E1 is expected to be available in the U.S. in the second quarter of this year, while it will be available this March in Europe. It’s also compatible with the Translator V4 for users who wish to use both devices.
Stomski explained that AI is used to help identify human voices to filter out background noise that could otherwise interfere with the live translation or the transcript generated by the devices.
"All of the translations there are connections with AI of course, but starting with the earbud – here we’ve got a model that helps us to recognize the human voice because we’re checking what is going into the microphone and we’re checking with the AI model if this is a human voice or this is like a car or a dog," he said.
The Vasco V4 Translator. (Courtesy of Vasco Electronics / Fox News)
GET FOX BUSINESS ON THE GO BY CLICKING HERE
Stomski added that the device uses several models from different providers, in addition to in-house solutions to cross-check speech translations as well as the image-to-text translations that Vasco’s tools can provide.
"We are constantly checking the quality of the translation for every pair because we cannot do it in a common language," he explained. Those translations are sent to Vasco’s cloud, which helps semi-automatically test the roughly 6,000 pairs of languages covered by the devices before those results are checked by humans.
The Gift of Words: Rotary Club donates dictionaries - The Post and Courier - Dictionary
Members of the Kingstree Rotary Club recently distributed dictionaries to third grade students at Greeleyville Primary School, Kenneth Gardner Leadership Academy, and Hemingway Elementary School as part of The Rotary Dictionary Project, nationwide effort to provide young students with their own dictionaries.
Photo Provided
A group of third graders received a special gift recently—the gift of words. The Rotary Club of Kingstree has continued its ongoing charitable donations in support of education with its annual donation of Merriam-Webster dictionaries to third grade students at Greeleyville Primary School, Kenneth Gardner Leadership Academy, and Hemingway Elementary School.
Photo Provided
The Rotary Dictionary Project is a nationwide effort to provide young students with their own dictionaries. Members of the Kingstree Rotary Club distributed the dictionaries to Williamsburg County School District third grade students the week of January 16.
Rotary members brought joy to the children’s faces when they learned the book was theirs to keep making the project worthwhile.
In Free Florida, the Dictionary Is Dangerous to Your Children - FlaglerLive.com - Dictionary
Book bans are right up there with censorship, the desecration of cultural artifacts and the whitewashing of history, all of which Florida now does routinely in the name of development and white nationalist purity. But I admit: I had a lot of fun reading over at least parts of the list of 1,600 books the Escambia County school system has removed from shelves, supposedly to review and potentially ban.
Librarians review books as a matter of course. They are professionals trained in the art of calibrating the right books to the right school audiences. So there’s nothing wrong with reviewing books along those lines. The problem is that those books have already gone through that process. Most have been on the shelves for years. Many are classics, some are dictionaries–dictionaries–and one of them is the Guinness Book of World Records.
I don’t know what could possibly be objectionable in the Guinness book. Maybe the pitchforks of Escambia thought the book encourages children to drink dark beer. Apparently the book does give some attention to animal species, not human unfortunately, that can copulate “more than 50 times in the same three to four hours, all with the same female,” or female chimpanzees copulating with eight different males in 15 minutes. But I’m not sure how that’s sexually explicit. To me it reads more like the set up to one of those brain-twisting math questions on the SAT.
Among the less esoteric rejections, we see titles by Walt Whitman, Sandra Day O’Connor Thurgood Marshall, just about every book by Maya Angelou, even though that famous passage of her getting raped when she was 9, in I Know Why the Cage Bird Sings, is all of two lines, rendered in metaphor: “Then there was the pain. A breaking and entering when even the senses are torn apart.” And that’s it.
Many titles that have been making headlines for the past two years are on an earlier Escambia list that led to a federal lawsuit. But that’s old news. Silencing Ann Frank’s Diary isn’t. Now we’re into not just censorship, but the erasure of people and memories, the erasure of man’s inhumanity toward particular people and races, as is the case with the removal of books by William Faulkner, Alice Walker and Toni Morison.
By the time you find that two books by Gabriel Garcia Marquez, the greatest storyteller of the last century, have been removed, it’s no longer surprising that the Encyclopedia of World Costumes is also on the ban list. No doubt, Escambia schools think that children could become gay by spending time looking at frou frous. It’s absurd enough for Ripley’s Believe It or Not, except that that title is banned in Escambia, too.
You can argue that book bans in the age of Amazon, Google and universal libraries are irrelevant. To some extent that’s true. Those of us with means can acquire any book we please, often overnight. But books aren’t cheap, and the days of online bargains are pretty much over. Besides: Americans don’t read much anymore and don’t even know what their children should read. So for a big segment of our school population, school libraries are it. Students either get their books there, or they don’t get them. Their cultural literacy is disproportionately at the mercy of what librarians put on shelves–what librarians are allowed to put on shelves.
Most reading discoveries are by browsing–the serendipity of finding a great book, discovering a writer who seems to speak to you personally, a story that makes you feel less alone, less of the freak everyone else makes you think you are, more of the human being that you have always been but are afraid to acknowledge. Those are the books that can make a life-changing–a life-affirming–difference in children’s lives. Those are the books that are being removed.
A few people who call themselves parents but are really frustrated bullies who want everyone else to lead the miserable lives they do, at least when they’re not engaging in threesomes, have successfully made black holes of Florida’s school and classroom libraries and further marginalized slews of children whose one solace might have been that one book.
It may not have been any one of the 23 Stephen King books banned, or the Grisham and Crichton and Koontz and maybe even the Picoults books that are literature’s equivalent of that gluey orange sauce McDonalds slathers on its burger imitations. But who are we to say what strikes a chord with a child’s imagination, what speaks to a child’s sense of wonder and self-discovery? Right now in Florida we may no longer ask the question. The bans have it. The rest is irrelevant.
Orange County Public Schools spent $400,000 in tax dollars in overtime pay to media specialists to draw up their own list of nearly 700 titles now found unacceptable (including Milton’s Paradise Lost and the one regained, Saul Bellow’s Herzog, and of course Joseph Heller’s Catch-22, now a metaphor for public schools, but also, thank heavens, Ayn Rand’s Fountainhead: no one will miss her felony-worthy prose). Public money, spent to silence history and deny students literary fun or discovery on the bogus, never-tested assumption that these books harm their readers.
Flagler County went through its own round of deshelving libraries, though it’s done so in more pernicious ways. We don’t really know how many books have been banned because most have been “weeded.” That’s a deceptive term that pretends that removing a book because it’s not read as much or may have been bought by mistake is not book-banning. It is. Teachers and librarians would rather err on the side of caution, otherwise they’d get sued or fired. So you can be certain that whatever happened in Escambia is happening in Flagler and across the state.
And that’s what they call Free Florida.
Pierre Tristam is the editor of FlaglerLive. A version of this piece airs on WNZF.
In Free Florida, the Dictionary Is Dangerous to Your Children - FlaglerLive.com - Dictionary
Book bans are right up there with censorship, the desecration of cultural artifacts and the whitewashing of history, all of which Florida now does routinely in the name of development and white nationalist purity. But I admit: I had a lot of fun reading over at least parts of the list of 1,600 books the Escambia County school system has removed from shelves, supposedly to review and potentially ban.
Librarians review books as a matter of course. They are professionals trained in the art of calibrating the right books to the right school audiences. So there’s nothing wrong with reviewing books along those lines. The problem is that those books have already gone through that process. Most have been on the shelves for years. Many are classics, some are dictionaries–dictionaries–and one of them is the Guinness Book of World Records.
I don’t know what could possibly be objectionable in the Guinness book. Maybe the pitchforks of Escambia thought the book encourages children to drink dark beer. Apparently the book does give some attention to animal species, not human unfortunately, that can copulate “more than 50 times in the same three to four hours, all with the same female,” or female chimpanzees copulating with eight different males in 15 minutes. But I’m not sure how that’s sexually explicit. To me it reads more like the set up to one of those brain-twisting math questions on the SAT.
Among the less esoteric rejections, we see titles by Walt Whitman, Sandra Day O’Connor Thurgood Marshall, just about every book by Maya Angelou, even though that famous passage of her getting raped when she was 9, in I Know Why the Cage Bird Sings, is all of two lines, rendered in metaphor: “Then there was the pain. A breaking and entering when even the senses are torn apart.” And that’s it.
Many titles that have been making headlines for the past two years are on an earlier Escambia list that led to a federal lawsuit. But that’s old news. Silencing Ann Frank’s Diary isn’t. Now we’re into not just censorship, but the erasure of people and memories, the erasure of man’s inhumanity toward particular people and races, as is the case with the removal of books by William Faulkner, Alice Walker and Toni Morison.
By the time you find that two books by Gabriel Garcia Marquez, the greatest storyteller of the last century, have been removed, it’s no longer surprising that the Encyclopedia of World Costumes is also on the ban list. No doubt, Escambia schools think that children could become gay by spending time looking at frou frous. It’s absurd enough for Ripley’s Believe It or Not, except that that title is banned in Escambia, too.
You can argue that book bans in the age of Amazon, Google and universal libraries are irrelevant. To some extent that’s true. Those of us with means can acquire any book we please, often overnight. But books aren’t cheap, and the days of online bargains are pretty much over. Besides: Americans don’t read much anymore and don’t even know what their children should read. So for a big segment of our school population, school libraries are it. Students either get their books there, or they don’t get them. Their cultural literacy is disproportionately at the mercy of what librarians put on shelves–what librarians are allowed to put on shelves.
Most reading discoveries are by browsing–the serendipity of finding a great book, discovering a writer who seems to speak to you personally, a story that makes you feel less alone, less of the freak everyone else makes you think you are, more of the human being that you have always been but are afraid to acknowledge. Those are the books that can make a life-changing–a life-affirming–difference in children’s lives. Those are the books that are being removed.
A few people who call themselves parents but are really frustrated bullies who want everyone else to lead the miserable lives they do, at least when they’re not engaging in threesomes, have successfully made black holes of Florida’s school and classroom libraries and further marginalized slews of children whose one solace might have been that one book.
It may not have been any one of the 23 Stephen King books banned, or the Grisham and Crichton and Koontz and maybe even the Picoults books that are literature’s equivalent of that gluey orange sauce McDonalds slathers on its burger imitations. But who are we to say what strikes a chord with a child’s imagination, what speaks to a child’s sense of wonder and self-discovery? Right now in Florida we may no longer ask the question. The bans have it. The rest is irrelevant.
Orange County Public Schools spent $400,000 in tax dollars in overtime pay to media specialists to draw up their own list of nearly 700 titles now found unacceptable (including Milton’s Paradise Lost and the one regained, Saul Bellow’s Herzog, and of course Joseph Heller’s Catch-22, now a metaphor for public schools, but also, thank heavens, Ayn Rand’s Fountainhead: no one will miss her felony-worthy prose). Public money, spent to silence history and deny students literary fun or discovery on the bogus, never-tested assumption that these books harm their readers.
Flagler County went through its own round of deshelving libraries, though it’s done so in more pernicious ways. We don’t really know how many books have been banned because most have been “weeded.” That’s a deceptive term that pretends that removing a book because it’s not read as much or may have been bought by mistake is not book-banning. It is. Teachers and librarians would rather err on the side of caution, otherwise they’d get sued or fired. So you can be certain that whatever happened in Escambia is happening in Flagler and across the state.
And that’s what they call Free Florida.
Pierre Tristam is the editor of FlaglerLive. A version of this piece airs on WNZF.
Subscribe to:
Posts (Atom)