Markus Freitag’s Post

✒ Could you explain the difference between Google Translator and a translation made with GPT models? In short: 👉𝐂𝐨𝐧𝐭𝐞𝐱𝐭: Google Translator focuses on individual words or phrases, while GPT considers the entire text, capturing the global context more effectively. 👉𝐋𝐞𝐱𝐢𝐜𝐚𝐥 𝐀𝐜𝐜𝐮𝐫𝐚𝐜𝐲: Google Translator often provides more literal translations, staying faithful to the original words. GPT, on the other hand, interprets the overall meaning, delivering translations that are more fluid and natural. In Practice: GPT interprets the whole message, adapting it better to the context. It can offer a more nuanced translation of idiomatic expressions or cultural references. 👇 Watch the video below! And you? How have you used GPT for translation tasks, or are you still relying on Google Translator? #gpt4o #azureopenai #clusterreply #innovation #azureaistudio

🌐 Concerning yet unsurprising news about the veracity of the internet! 🌐 📋 A recent arXiv paper reveals the shocking extent of AI-generated translations online. Out of 6.38 billion sentences studied, a whopping 2.19 billion were actually low-quality translations of English originals, often created by AI. 🤖💬 👩🏫 This has serious implications for anyone who may be exposed to inaccurate, misleading, or even harmful content online. How can we tell if a web page is machine translated or not? 🔍 The researchers suggest that one way to spot machine translated content is to look for multi-way sentences, which are sentences that can be found translated into several different languages. Easier said than done for the average user! These sentences tend to be shorter, more predictable, & less informative than human-written ones. They also come from low-quality sources, such as articles on how to be happy or tips for new boat owners. 🙅♀️ So, next time you’re browsing the web, particularly if it’s in a language other than English, be careful what you read & trust. Don’t let the machines fool you with their cheap tricks. 🙋♀️ A big shoutout to flesh & blood translators! 🤔 Have you spotted any inaccuracies due to how something's been translated...? https://lnkd.in/ehsaznfK #ELT #MachineTranslation #LanguageLearning #WebLiteracy #CriticalThinking  

Transformer Impact: Has Machine Translation Been Solved? - Google recently announced their release of 110 new languages on Google Translate as part of their 1000 languages initiative launched in 2022. In 2022, at the start they added 24 languages. With the latest 110 more, it’s now 243 languages. This quick expansion was possible thanks to the Zero-Shot Machine Translation, a technology where machine learning models learn to translate into another language without prior examples. But in the future we will see together if this advancement can be the ultimate solution to the challenge of machine translation, and in the meanwhile we can explore the ways it can happen. […] - https://lnkd.in/gkVzpt3F

Transformer Impact: Has Machine Translation Been Solved? - Google recently announced their release of 110 new languages on Google Translate as part of their 1000 languages initiative launched in 2022. In 2022, at the start they added 24 languages. With the latest 110 more, it’s now 243 languages. This quick expansion was possible thanks to the Zero-Shot Machine Translation, a technology where machine learning models learn to translate into another language without prior examples. But in the future we will see together if this advancement can be the ultimate solution to the challenge of machine translation, and in the meanwhile we can explore the ways it can happen. […] - https://lnkd.in/gkVzpt3F

Slator writes about a new open-source approach named GenTranslate to get better translations https://lnkd.in/e7ru2DQa. #machinetranslation #generatieveai #artificialintelligence #slator

"Like Darmok & Jalad, the memespace bubbles create epistemic black holes, internet niche-cliques shepherded by algorithm, (those freelancers who can code-switch between the memetic argot are called "translators".)" GPT: This sentence is rich with metaphors and concepts, particularly from the realms of speculative fiction and communication theory. Here’s a clearer and more detailed explanation: In the phrase, 'Like Darmok & Jalad, the memespace bubbles create epistemic black holes, internet niche-cliques shepherded by algorithm, (those who can code-switch between the memetic argot are called 'translators'.)', we encounter a multi-level analogy and a description of societal segmentation driven by technology, highlighting the underlying crisis of technological schism. The reference to 'Darmok & Jalad' alludes to a famous episode of 'Star Trek: The Next Generation,' where communication occurs solely through metaphor and allegory—people speak using historical and cultural references understandable only to those within the same community. This analogy is employed to illustrate how 'memespace bubbles' function—these are akin to self-contained ecosystems on the internet where specific jargon, memes, and cultural references form a unique language that may be incomprehensible to outsiders. These 'memespace bubbles' are described as creating 'epistemic black holes.' In this context, an 'epistemic black hole' refers to a scenario where certain knowledge or understanding is isolated within a specific group, with little to no information escaping to the wider world. This isolation makes it difficult for outsiders to understand or engage with the group, effectively preventing the exchange of information across different societal segments, leading to fragmentation. The bubbles are 'shepherded by algorithm,' meaning algorithms—likely through social media platforms and search engines—curate and reinforce these niche cliques by continuously feeding them content that aligns with their established beliefs and knowledge. This algorithmic curation intensifies the boundaries around these groups, making them more insular until they are entirely cut off from one another. Finally, individuals capable of 'code-switching between the memetic argot' are termed 'translators.' In this context, 'memetic argot' refers to the specialized language and symbols used within each memespace bubble. Translators are adept at navigating multiple memespaces, understanding and communicating across these diverse cultural and linguistic islands. They play a pivotal role in bridging divides in a fragmented digital society by conveying and interpreting meanings across different internet subcultures. This scenario paints a picture of a future where digital communication leads to highly specialized communities that are often isolated from each other, with only a select few capable of spanning these divides for those who can afford their services.- Jamie Curcio

Elevate your translation quality with LILT's Verified Translation workflow, combining the power of human intelligence and real-time multilingual Large Language Model fine-tuning. Say goodbye to quality issues highlighted in Amazon's recent report on machine-translated web content. Read the report here: https://lnkd.in/gHmfDT6G #TranslationQuality #LILT #LanguageTechnology #VerifiedTranslation #GenerativeAI #EnterpriseAI

Elevate your translation quality with LILT's Verified Translation workflow, combining the power of human intelligence and real-time multilingual Large Language Model fine-tuning. Say goodbye to quality issues highlighted in Amazon's recent report on machine-translated web content. Read the report here: https://lnkd.in/g8eghK3t #TranslationQuality #LILT #LanguageTechnology #VerifiedTranslation #GenerativeAI #EnterpriseAI

https://lnkd.in/eyrdyd-W With the rising demand for accessible and engaging video content across global markets, Video Translation offers a seamless solution to overcome language barriers. This launch includes an Azure Speech for customers to try out with their own video assets, with turn-key capabilities such as: Dialogue extraction and translated subtitles generation GPT reformulation with improved translation quality and automatic time alignment Prebuilt neural voices with content editing for precise alignment and translation preference manually Using the personal voice capability (will be available with limited access restrictions)

Let me depict a scenario about how tools might be in the future. Suppose you found a neat tool, a simple language translator, that you like and find useful. But you miss something, something you’d like to change or add. You’re a bit disappointed because it would be perfect if only it had that missing feature. You might ask the original developer to add it, but, you know, things are not always so easy. You end up using it, settling for it as it is, or continuing your search for the perfect tool... …or… you can click a URL that brings you to a “tool editor” where you can add or change the feature you’re looking for using natural language, creating your own variation, just for you or your team. Well, it might be the present. You can experience this latter scenario with our first micro-tool. It’s a simple translator tool that you can use or modify at your discretion. Here more information: https://lnkd.in/d3mA7RgJ