Kuba Krchak

Innovations in #artificialintelligence and #naturallanguageprocessing are expected to bridge existing gaps, particularly in domains like skill taxonomies and digital twin data sharing. The Semantic Web is poised to transform how we interact with data on the Internet, offering new opportunities for innovation and efficiency. Understand what is Semantic Web, What makes the difference in “Semantic Web”? The Evolving Vision of the Semantic Web, and more. https://lnkd.in/gM7Zu9WX #semanticweb #semanticatechnology #innovation #digitaltwins #AI #latestarticle Nwachukwu Ebenezer

2

A very nice article expalining the Small Language Models. https://lnkd.in/dgTuspiS The Microsoft Phi-3 is one of these.

3

From simple LLMs to a future where large parts of the enterprise are controlled by MAS 🏢 Prof. Dr. August-Wilhelm Scheer, Founder and Sole Shareholder of Scheer Companies, and Dr. Wolfram Jost, Managing Director and CTO, share their insights on the evolution of AI in the enterprise. They discuss the journey so far and predict where AI is headed. 📰 Check out the article to learn more: https://lnkd.in/eBZehaj8 Plus, get a sneak peek into the future of Scheer PAS and discover how our Process iPaaS approach can help you navigate the ever-changing business climate. What do you think about the future of AI in enterprise? #AI #ScheerPAS #MAS #Enterprise #iPaaS

7

Use #LLM to improve the quality of #classification in your #knowledgegraph. Evaluated over #Wikidata and #CaLiGraph. By Paul Groth and Bradley Allen https://lnkd.in/eKEZWM4y #ESWC2024 #ESWC #linkeddata #semanticweb #AI #classhierarchy

26

#AI Solution #Architect Journey (#ELVTR). 4th Week, 3rd Cohort: #ML deployment process parallel to Petra discovery. There is something inspiring about the moment when the #AI and #GenAI solution you have created is deployed. It is like a recent discovery in Petra, a hidden tomb unearthed after centuries of mystery. Archaeologists found a skeleton holding a vessel closely resembling Indiana Jones's 'Holy Grail.' Our AI deployment process is not that exciting, but the level of satisfaction is comparable. We're in the 4th week of the AI Solution #Architect course by #ELVTR, and it's all about bringing the ML creations to life. This week, we embarked on the process of ML deployment. Only our hidden tomb was an #NLP Text Toxicity Moderation app, and the treasure was its deployment on #HuggingFace, a real-world-impact website. It is the kind of app that stands between a healthy online conversation and the chaos of the toxic social media landscape. As we navigated the deployment process, we tackled the requirements of hardware and platform deployment. It's a balancing act, often less about perfect answers and more about practical trade-offs. Like archaeologists weighing whether to excavate immediately or stabilize the surroundings first, we, too, had to decide how best to bring our ML app into the world. The ML #pipeline is not just about shipping code; it's about delivering reliability and predictability. Just as we had to carefully extract a vessel from Petra's tomb without disturbing the fragile remains, we had to ensure that each step in our ML pipeline was deliberate and precise. After deployment, the story doesn't end. AI #analytics, ML #feedback loops, and update cycles come into play, creating a constant growth cycle. We must monitor our model once it's out in the wild, gather feedback, and iterate to align it with user expectations. Deploying is not the end; it's the beginning of real learning. Just like an ancient artifact starting its journey of study and understanding in the hands of experts, our ML model begins its true life once deployed. Week 4 of this course reminded us that AI isn't just a collection of algorithms; it's about what you build with it, where you put it, and how you ensure it grows more intelligent with every use. I can't wait to see what's in store in our continuing AI journey. Class [& 2025 class]: https://lnkd.in/gbcB_t7k App #HuggingFace: https://lnkd.in/g2kB6sEJ This article complies with The First Law of AI Collaboration: https://lnkd.in/gVDfgq4R Please #repost, #follow, and #thumbs-up. I hope you enjoy reading it as much as I enjoy writing it. #innovation, #hollygrail #DucHaba

8

💥💥💥 Thinking LLMs: General Instruction Following with Thought Generation | 1Meta FAIR, University of California, Berkeley, New York University Tianhao Wu, Janice Lan, Weizhe Yuan, Jiantao Jiao, Jason Weston, Sainbayar Sukhbaatar Abstract LLMs are typically trained to answer user questions or follow instructions similarly to how human experts respond. However, in the standard alignment framework they lack the basic ability of explicit thinking before answering. Thinking is important for complex questions that require reasoning and planning -- but can be applied to any task. We propose a training method for equipping existing LLMs with such thinking abilities for general instruction following without use of additional human data. We achieve this by an iterative search and optimization procedure that explores the space of possible thought generations, allowing the model to learn how to think without direct supervision. For each instruction, the thought candidates are scored using a judge model to evaluate their responses only, and then optimized via preference optimization. We show that this procedure leads to superior performance on AlpacaEval and Arena-Hard, and shows gains from thinking on non-reasoning categories such as marketing, health and general knowledge, in addition to more traditional reasoning & problem-solving tasks. 👉 https://lnkd.in/d2sBPFJm #machinelearning

1

🧠👀 The legendary Vinod Khosla of Khosla Ventures just published a set of "Plausible Tomorrows (2045-2049)". #1 : expertise will be free . We will be capable of having accessible AI-based near free tutors for every child, and doctors for every citizen 24x7 #2 : labor will be near free . We will have a billion bipedal (and other) robots freeing humans from the servitude of undesirable jobs, creating a larger industry than the auto industry #3: computer use will grow expansively #4: AI will play a large role in entertainment & design #5: internet access will be mostly by agents (agents are the future workers, in my book "AGENTWARE") More in the slides... it's hard for us humans to envision exponential technologies. Mr. Khosla has experience in investing in the long term. 🧠 What is your view on these plausible tomorrows? Do you agree with them all? #plausibletomorrows #ai

9

2 komentáře

The future of Medicine powered by AI might be fantastic. We will talk about some aspects of it at the next AI education organised by Hrvatska gospodarska komora (Tajana Kesić Šapić and Iva Nappholz). Here is one warmup article about the possibilities - of AI hospitals in China: https://lnkd.in/draKkJED Wonder what the European answer will be ... #ai #health #future #education Vedran Antoljak

21

16 komentářů

This will unlock vector search for the 10-100x number of use cases where it has been too expensive until now. A must read if you're using embeddings.

6

🥦 Broccoli AI (coined by Demetrios Brinkmann): AI that doesn't necessarily taste good for the organization (in terms of playing into the hype), but that is needed. We had Bengsoon Chuah on the Practical AI Podcast [link in the comments], who is a data scientist practicing Broccoli AI at its best. This is a practical story of a practitioner actually getting AI into production in a challenging, enterprise environment! #datascience #AI #BroccoliAI #podcast

32

7 komentářů

Where Do You Still Need Machine Learning? (Guest: Srujana Kaddevarmuth) https://buff.ly/4eWRj4a via Andreas Welsch of Intelligence Briefing on Thinkers360 #AI #DigitalTransformation #GenerativeAI

3

📚Good read!📚 Eyal Trabelsi shares a comprehensive testing strategy for #machinelearning systems. As he states, testing ML is hard because it's "probabilistic by nature, and must account for diverse and dynamic real-world conditions." In this article, Eyal shares the foundation for the stages that will help you overcome these challenges : 1. 𝗖𝗜 𝗣𝗶𝗽𝗲𝗹𝗶𝗻𝗲 🔄 : Automate the testing process to ensure consistency and efficiency. 2. 📝 𝗦𝘆𝗻𝘁𝗮𝘅 𝗧𝗲𝘀𝘁𝗶𝗻𝗴: Validate the code to catch syntax errors early in the development cycle. 3. 𝗗𝗮𝘁𝗮 𝗖𝗿𝗲𝗮𝘁𝗶𝗼𝗻 𝗧𝗲𝘀𝘁𝗶𝗻𝗴 👌: Verify the data generation process to maintain data quality and integrity. 4. 🦾 𝗠𝗼𝗱𝗲𝗹 𝗖𝗿𝗲𝗮𝘁𝗶𝗼𝗻 𝗧𝗲𝘀𝘁𝗶𝗻𝗴: Evaluate model performance using various metrics to ensure accuracy and reliability. 5. 𝗘𝟮𝗘 𝗧𝗲𝘀𝘁𝗶𝗻𝗴 ♾️: Conduct end-to-end tests to simulate real-world scenarios, ensuring the entire ML pipeline works seamlessly. 6. 📦 𝗔𝗿𝘁𝗶𝗳𝗮𝗰𝘁 𝗧𝗲𝘀𝘁𝗶𝗻𝗴: Test the final artifacts to confirm their readiness for production deployment. Read the very detailed, in depth article here https://lnkd.in/dXMbArkt

7

Ramblr’s Data Engine now supports automated #SceneGraph generation! This powerful feature bridges the gap between visual scenes and human language, enabling advanced vision-language tasks and seamlessly injecting video context into #LLM prompts. Drago Guggiana Nilo walks you through the functionality and shows you how to get started. Get your free account at https://ramblr.ai and try automated #SceneGraph generation on Ramblr.ai®'s platform Stay tuned for more - We are constantly improving our  #DataEngine enabling advanced #VideoUnderstanding. #AR #MR #VideoIntelligence #ContexualAI #AIforThePhysicalWorld 

10

Let’s dwell deeper into understanding how #RAG works through a hands-on implementation using #SingleStore as a #vectordatabase to store the vector data. https://lnkd.in/ehTXieXZ

2

#SyntheticData for #LLMs or #Synthetic #Text #Generation Make your #GenerativeAI projects have high-quality outcomes! Enhance LLMs with improved data privacy and augmentation. Join #Fabric today at ydata.ai/register and leave a comment for early access! #YData #AI #genai #textdata #LLM #NLP #DeepLearning

10

5 komentářů

Check out how the world’s first #generativeAI models for #OpenUSD development will be available as NVIDIA NIM microservices.

6

The Rails 8 feature set perfectly complements SQLite's power in creating resilient, high-performance production apps, but still the question lingers: Can I really go all-in on #SQLite? Join Stephen Margheim at #RailsWorld as he illustrates how to leverage Rails and SQLite's full potential in your next venture, and when SQLite does and doesn't make sense for your application. https://lnkd.in/eaBstScy

52

1 komentář