#AI Language Translation

[ailifehacks]

AI for Language Translation: How GPT and BERT Are Changing Multilingual Communication

Explore how AI models like GPT and BERT revolutionize language translation, enabling seamless multilingual communication across the globe. In an increasingly globalized world, effective communication across language barriers is essential. Artificial Intelligence (AI) has emerged as a transformative force in language translation, with models like GPT (Generative Pre-trained Transformer) and BERT…

[prokopetz]

It's no secret that Google Translate's results have been getting steadily worse ever since they switched to LLM-based dictionaries, but a new thing is whenever I click "Detect Language" it now automatically sets my target language to Spanish? I'm from Saskatchewan.

[14dayswithyou]

Olivia fell asleep in one of the aisles again...

how do you feel about ren being used in an ai chatbot advertisement?

⌞♥⌝ Thank you for bringing this to my attention. I genuinely despise everything about this so much, and if anyone comes across this ad on TikTok (or anywhere else), I'd really appreciate it if you could report the video and not give it any further engagement.

I'm vehemently against the use of AI that negatively impacts artists, writers/authors, developers, creators, etc., and I don't condone the use of my art and IP without my knowledge and explicit consent — especially when it's being used in a paid ad/sponsorship. It's extremely disrespectful and I have no respect for those who do it.

[reasonsforhope]

"As a Deaf man, Adam Munder has long been advocating for communication rights in a world that chiefly caters to hearing people. 

The Intel software engineer and his wife — who is also Deaf — are often unable to use American Sign Language in daily interactions, instead defaulting to texting on a smartphone or passing a pen and paper back and forth with service workers, teachers, and lawyers. 

It can make simple tasks, like ordering coffee, more complicated than it should be. 

But there are life events that hold greater weight than a cup of coffee. 

Recently, Munder and his wife took their daughter in for a doctor’s appointment — and no interpreter was available. 

To their surprise, their doctor said: “It’s alright, we’ll just have your daughter interpret for you!” ...

That day at the doctor’s office came at the heels of a thousand frustrating interactions and miscommunications — and Munder is not isolated in his experience.

“Where I live in Arizona, there are more than 1.1 million individuals with a hearing loss,” Munder said, “and only about 400 licensed interpreters.”

In addition to being hard to find, interpreters are expensive. And texting and writing aren’t always practical options — they leave out the emotion, detail, and nuance of a spoken conversation. 

ASL is a rich, complex language with its own grammar and culture; a subtle change in speed, direction, facial expression, or gesture can completely change the meaning and tone of a sign. 

“Writing back and forth on paper and pen or using a smartphone to text is not equivalent to American Sign Language,” Munder emphasized. “The details and nuance that make us human are lost in both our personal and business conversations.”

His solution? An AI-powered platform called Omnibridge. 

“My team has established this bridge between the Deaf world and the hearing world, bringing these worlds together without forcing one to adapt to the other,” Munder said. 

Trained on thousands of signs, Omnibridge is engineered to transcribe spoken English and interpret sign language on screen in seconds...

“Our dream is that the technology will be available to everyone, everywhere,” Munder said. “I feel like three to four years from now, we're going to have an app on a phone. Our team has already started working on a cloud-based product, and we're hoping that will be an easy switch from cloud to mobile to an app.” ...

At its heart, Omnibridge is a testament to the positive capabilities of artificial intelligence. "

-via GoodGoodGood, October 25, 2024. More info below the cut!

To test an alpha version of his invention, Munder welcomed TED associate Hasiba Haq on stage. 

“I want to show you how this could have changed my interaction at the doctor appointment, had this been available,” Munder said. 

He went on to explain that the software would generate a bi-directional conversation, in which Munder’s signs would appear as blue text and spoken word would appear in gray. 

At first, there was a brief hiccup on the TED stage. Haq, who was standing in as the doctor’s office receptionist, spoke — but the screen remained blank. 

“I don’t believe this; this is the first time that AI has ever failed,” Munder joked, getting a big laugh from the crowd. “Thanks for your patience.”

After a quick reboot, they rolled with the punches and tried again.

Haq asked: “Hi, how’s it going?” 

Her words popped up in blue. 

Munder signed in reply: “I am good.” 

His response popped up in gray. 

Back and forth, they recreated the scene from the doctor’s office. But this time Munder retained his autonomy, and no one suggested a 7-year-old should play interpreter. 

Munder’s TED debut and tech demonstration didn’t happen overnight — the engineer has been working on Omnibridge for over a decade. 

“It takes a lot to build something like this,” Munder told Good Good Good in an exclusive interview, communicating with our team in ASL. “It couldn't just be one or two people. It takes a large team, a lot of resources, millions and millions of dollars to work on a project like this.” 

After five years of pitching and research, Intel handpicked Munder’s team for a specialty training program. It was through that backing that Omnibridge began to truly take shape...

“Our dream is that the technology will be available to everyone, everywhere,” Munder said. “I feel like three to four years from now, we're going to have an app on a phone. Our team has already started working on a cloud-based product, and we're hoping that will be an easy switch from cloud to mobile to an app.” 

In order to achieve that dream — of transposing their technology to a smartphone — Munder and his team have to play a bit of a waiting game. Today, their platform necessitates building the technology on a PC, with an AI engine. 

“A lot of things don't have those AI PC types of chips,” Munder explained. “But as the technology evolves, we expect that smartphones will start to include AI engines. They'll start to include the capability in processing within smartphones. It will take time for the technology to catch up to it, and it probably won't need the power that we're requiring right now on a PC.” 

At its heart, Omnibridge is a testament to the positive capabilities of artificial intelligence. 

But it is more than a transcription service — it allows people to have face-to-face conversations with each other. There’s a world of difference between passing around a phone or pen and paper and looking someone in the eyes when you speak to them. 

It also allows Deaf people to speak ASL directly, without doing the mental gymnastics of translating their words into English.

“For me, English is my second language,” Munder told Good Good Good. “So when I write in English, I have to think: How am I going to adjust the words? How am I going to write it just right so somebody can understand me? It takes me some time and effort, and it's hard for me to express myself actually in doing that. This technology allows someone to be able to express themselves in their native language.” 

Ultimately, Munder said that Omnibridge is about “bringing humanity back” to these conversations. 

“We’re changing the world through the power of AI, not just revolutionizing technology, but enhancing that human connection,” Munder said at the end of his TED Talk. 

“It’s two languages,” he concluded, “signed and spoken, in one seamless conversation.”"

-via GoodGoodGood, October 25, 2024

[frostwinddandelion]

[selfloathingextraordinaire]

With what's happened with Duolingo, I'm looking for a new language app, but need some help.

Is there a Japanese language app that is:

Great for people with learning disabilities (my brain struggles with language and has to learn it very slowly)

Kid friendly

Colorful

I've seen a bunch of posts with all these recommendations, but it's super overwhelming and I have no idea where to start.

[tardisscans]

Chinen Hidekazu x Hino Yusuke - A slightly relaxed Kamen Rider BBQ two-person talk

I'm starting to learn Japanese and I’ve been into Toku lately, so I decided to give it a go and upload a rough translation of the TTFC BBQ the main cast had a bit ago.

Sadly, Tumblr is too wonky to upload any kind of video over 5 minutes, so I used GoFile (The file deletes itself after a week tho TT)

[ghostjelliess]

Had to get serious with a .3 mm pen cus the .5 jelly rolls smudged too much 😭 now I get why Muji pens are like that.

Learning essential phrases, thanks Duo! 🙏

Just curious how it translates! It worked!

Oh, I wonder how my worksheets translate. It will look funny, like: soup soup soup soup soup soup...

Ope... 👀 Should have stayed curious. 😒

[casualavocados]

Are you here to pick me up? Going to the bar or the gang? The bar.

KISEKI: DEAR TO ME Ep. 12 + BONUS:

[mawrrbid]

Touchstarved incorrect quotes as shit I and my friends said (without any context) - pt.1 probably

pt.2 - pt.3

✦ ✧ ✦ ✧ ✦

Leander: If you think about it, beer is technically soup. :)

--

MC: You're a cutie patootie. Vere: Wrong, I'm a slutty putati.

--

Kuras: There's many things I wanna say but I am not sure church will appreciate it. MC: OOGA BOOGA.

--

Mhin: I'm a spilled cup of expresso, wasted and bitter.

--

MC: Why are you on all four on the ground like a slut? Ais: I AM LOOKING FOR SOMETHING TO SMOKE!

[malinowaj]

coworker asking me why i translated a set of instructions myself and didn’t just use chatGPT. i’ve never wanted to say ”fuck you” to someone more than i did then.

[tarn-ati0n]

What the FUCK is going on with youtube???

[skeletalheartattack]

Subtitles below the keep reading:

Hey you, shut your mouth and look at my paw! DON'T FORGET!! [Incomprehensible sped up gibberish] This... Journey... Money... Loads of coins. [Incomprehensible sped up gibberish] What-about-her? What-about-her? ... What-about-her? 'She still like me? [Quiet but mostly incomprehensible gibberish about subtitles] O P S O P N O-1 1. Here's the spell: Love the mermaid, for sure! The mermaid is HAPPY! Okay! It's pretty normal for a fish, right? Guuuyyyssss, beeeee caaarefuuulll wiiiiith theeee GIIIIIRRRRLSSS!!! [Incomprehensible] Oh! Silly! Oh yes! Lamb chop boy! [Incomprehensible] [Very quietly, while white noise is playing over it] Goood eevening, aand weeelcome too the shoooowww... [In the background] Ohhh, mooney!

[gayeilgeoir]

Honestly I don’t trust a single one of you fuckers that are so “anti ai” when it comes to school assignments because look me in the eyes and tell me without lying that you’ve never used Google translate for a language class.

“Oh but it was only for one word” yeah, and? You still harass the fuck out of people for using Chat GPT for one word, what’s the difference?

“But how else am I meant to figure out the word?” DICTIONARIES!!!! They’re actually really useful, and if you can only get one word at a time, you’ll actually learn how to formulate sentences in the language!

“It’s not that big of a deal” YES IT IS!!! Google translate produces shitty, grammatically incorrect translations. It is god awful at nuance, colloquialisms, non-standardised language, and is actively harming your ability to form sentences in that language and your fluency.

“I only use it for pronunciation guides” THATS STILL BAD! Again, a machine cannot teach you how to speak! If you put a child in front of an AI that spoke to it 24/7, it wouldn’t learn language properly, you’d be arrested for child abuse! So how do you expect that to happen for you?

This is not me defending the use of AI academically, it’s quite the opposite. I’m just beyond tired of peoples blatant disregard for how harmful things like google translate are, and not looking at it like a form of AI that ruins how you learn the language just as much as using Chat GPT ruins your ability to learn how to write essays, or write anything on a certain subject.

A machine cannot ever be even remotely as good at translation as a human, because languages are such an innately human thing.

[unicoherent]

[dedalvs]

can you make a translator for firish i want to use it in my rps i have with friends

I've actually gotten this question a couple times, which is great! But this type of thing just isn't possible with a conlang. It has nothing to do with the quality of the conlang or the level of completion (i.e. the amount of vocabulary, how much of the grammar has been recorded, etc.), and I'll tell you specifically why.

First, you may have seen "translators" for various languages online like LingoJam. LingoJam not only has translators for a bunch of different languages, but allows you to make your own translators. The way these work, though, is you write down a word in one language and write its translation into another—something like:

English > Spanish

I > yo

am > soy

to > a

the > el

store > tienda

going > yendo

That is, you put in one to one correspondences, and that's what it has to work with. Once you're done, if you ask for a translation, it looks up the words and sees what's available and it spits back what it has, in order. If we had this very minimal English to Spanish dictionary (which is 100% accurate, by the way! That is, all of these English words can be translated as all of these Spanish words), you could ask LingoJam to translate the following into Spanish...

I am going to the store.

...and you would get...

Yo soy yendo a el tienda.

Now, if you speak Spanish, you'll see all the places this went wrong. (Short version: You don't always need subjects pronouns in Spanish; you use a different helping verb for "to be x'ing" in Spanish; you rarely actually use this "to be x'ing" construction in Spanish; the present tense is sufficient; though el means "the", it's the wrong gender for tienda—analogous to saying "an store" as opposed to "a store" in English.) And you can actually avoid this in LingoJam by adding phrases on top of single words:

English > Spanish

the store > la tienda

I am going > voy

But you can imagine how much work that would be...

The reason why things like LingoJam are so popular, though, is because imagine if you knew nothing about Spanish. Typing in "I am going to the store" and having it instantly spit out "Yo soy yendo a el tienda" is pretty darn satisfying! If you don't know it's wrong but you're happy with it, what's the problem?

Now, a language like Spanish is huge, so it's easier to get accurate Spanish translations online than it is to get accurate Korean translations online—and it's easier to get accurate Korean translations online than accurate Tigrinya translations online, etc. The reason for that takes us to Google Translate.

I think most people know that with LingoJam, you get what you pay for. Google Translate, on the other hand, is much more sophisticated, and much more accurate. It's not 100%, but it's pretty darn good—for widely spoken languages. This is why.

Way back when, Syfy facilitated a chat between me and the folks at Google Translate because they wanted to see if Google and I could work together to create a translator for a couple of my Defiance languages at TED in 2013. After all, we had a full two weeks. We could bang something like that out in two weeks, right? (lol no)

I learned then how Google Translate works. Google Translate doesn't actually know anything about the specific grammar of a language—maybe a couple language specific tweaks, but it's not as if you can go under the hood and find a full grammar of Spanish that tells you when to use the subjunctive, what all the conjugations are, etc. Instead, what Google Translate has is a database (i.e. Google, along with Google Books, Google Scholar, etc.) with tons of, presumably, fluent documents written in the various target languages offered on Google Translate. They also have faithful translations of those documents—not all, but a percentage. Google Translate uses that information to predict what a given sentence in one language will turn into in another.

In order to do this successfully, Google Translate needs BILLIONS of documents to troll. And it has that. It has BILLIONS of articles written in Spanish and translated to English. That's why the English to Spanish translation is as good as it is.

Now, having said that, anyone who's bilingual in English and Spanish knows that Google Translate isn't perfect. Sometimes it's pretty good, but sometimes it produces a lot of clunky, unnatural, or even incorrect translations. This is because there isn't a human back there calling the shots.

But that's its best translator. Now imagine translating between English and Samoan (one of the other languages it offers). There are EXPONENTIALLY more online articles in Spanish than Samoan. Consequently, the translations you get between English and Samoan on Google Translate are absolutely no guarantee.

And bear in mind, there's a kind of minimum threshold they work with before adding a language to Google Translate. If Samoan is on there and not Fijian, it's because there's that much more Samoan online than Fijian.

Now let's go back to conlangs. What Google Translate wants is BILLIONS of articles written online in the target language. Forget how complete the grammar of a conlang is, whether you can find that description online, or how many thousands of words the conlang has. How many fluent articles are there written in that conlang that are online? How many can one person to? How about a team of people? And how many conlangs have that?

This is why Google Translate has Esperanto and nothing else. Esperanto has been around for 136 years, and in that time there have been a good number of people who have learned to speak it fluently, and have written things (poems, articles, books) that are now online. It is as much as Spanish? Certainly not, but it is enough to hit Google Translate's minimum threshold, and so it's available.

Assuming you have a conlang with a full grammar and a good amount of vocab, if it were popular, it might have enough available material for Google Translate to work with 125 years from now. But at the moment, it's not possible. That says nothing about the language: It's about how Google Translate works.

And bear in mind, Google Translate is, at the moment, our best non-human translator.

If predictive-AI gets good enough that it can learn the grammar of a language, then it may be possible to produce a translator for a new conlang. That, though, is not the goal of Google Translate. Maybe ChatGPT and things like it will get there one day, but even that isn't a dedicated language learning AI. We need an AI that doesn't work with billions of fluent articles, but works with two books: a complete grammar and a dictionary. If an AI can one day work with those two tiny (by comparison) resources and actually produce translations that are as good as or better than Google Translate, then we'll be at a "translation-on-demand" place that will be good enough to feed a new conlang to. At that point, it will simply be a matter of producing a grammar and lexicon of sufficient size for the AI to do its thing.

So, no, right now we can't do a Ts'íts'àsh translator. :( We can go over things like the sound system and basic grammar and you can create your own words to work with it... A lot more work, but hey, we don't have to churn our own butter or milk our own cows anymore! We've got time!