Matthias Fey

🚀 Exciting News! 🚀 We are thrilled to announce that Kumo AI's position paper "Relational Deep Learning: Graph Representation Learning on Relational Databases," has been accepted at #ICML 2024! 🎉 Join us in Vienna from July 21-27, where our very own Matthias Fey will be presenting this exciting new research area. 🔗 https://icml.cc/ In collaboration with Stanford University, we introduce Relational Deep Learning (#RDL), a new research area for applying deep learning on relational databases by viewing them as temporal, heterogeneous graphs. This method eliminates the need for manual feature engineering, leading to more accurate models built faster. 🚀 We’ve also developed #RelBench, a benchmark suite to facilitate research in this new area, using diverse datasets from Stack Exchange discussions to Amazon book reviews. 📚 Stay tuned for more updates on our presentation format, and check out our paper here: https://lnkd.in/dk24vbSg #ICML2024 #MachineLearning #AI #GraphNeuralNetworks #RelationalDeepLearning #DataScience #KumoAI #StanfordCollaboration

38

📅 On 07.11.2024, we invite you to the ScaDS.AI Dresden/Leipzig training "Big Data Processing on HPC" with Apurv Kulkarni and Wenyu Zhang. Apache Spark and Apache Flink are two typical #BigData analytics frameworks. Their APIs allow the development and testing of an application on a local workstation and later, without changing the source code of the application, distribute work to many computers when the local workstation is not sufficient anymore due to limited resources. The course focuses on the step from a local workstation to an #HPC environment and presents how the typical Big Data analysis workflow can be organized in an HPC environment. In this course participants will be introduced to running a data pipeline and data processing along with managing the configurations on the HPC environment, using Apache Flink and Apache Spark. Register here: 👉 https://lnkd.in/eYfmiMT8

15

Welcome to read the article "Integrated Gradients for Feature Assessment in Point Cloud-Based Data Sets" which is written by Mr. Markus Schwegler, Prof. Christoph Müller and Alexander Reiterer. This excellent article was published in issue 7 in 2023 and has been viewed 1313 times and cited 3 times! Article link: https://lnkd.in/gMCsed5z MDPI Fraunhofer IPM Furtwangen University Albert-Ludwigs-Universität Freiburg (Albert Ludwig University of Freiburg) #pointcloud #neuralnetwork #integratedgradients

2

Join us as we unpack the future of AI with Carsten Kraus in this follow-up article from Launch's #NavigatingAbroad series. Learn about the impactful role of small language models in making AI more accessible and the intricate dynamics of reinforcement learning. What implications will these technological advances have on the future landscape of AI? #AIUnpacked #FutureOfAI #LanguageModels #ReinforcementLearning #TechInnovation

6

🔍 Excited about Large Language Models (LLMs) revolutionizing data generation and labeling in Machine Learning? Join Ruben Bösche and Robin Suxdorf explore LLMs' power in their next upcoming session at MLCon! 📅 Date: Wednesday, June 26, 2024 🕓 Time: 16:00 - 16:45 #AI #DeepLearning #MachineLearning #LargeLanguageModels

9

For those who are interested in attending the Coding.Waterkant event but are unable or unwilling to commit to the full week, I have good news: the Monday evening session is open to everyone! Join us for a fascinating keynote speech "LAION and its Role in Open Source AI" by Christoph Schuhmann and take the opportunity to learn more about us and the unique Coding.Waterkant format. I encourage you all to mark your calendars and spread the word. I hope to see a strong turnout and many familiar faces there! #MachineLearning #AI #CodingWaterkant24 #opensource

14

1 Kommentar

🌟 Another SwissText Keynote revealed 🌟 We are thrilled to invite you to an insightful keynote titled "Attacks on large language models" by Leon Derczynski, affiliated with NVIDIA and ITU Copenhagen, at the Swiss Text Analytics Conference in Chur.     🔍 Abstract: Generative language models got a lot better recently; they escaped the research "yard" and now act in public. This has the consequence of risking reputational and other harms as models produce undesirable output. Simultaneously, generative language models have become alluring targets for hackers, who attempt to subvert them. Unlike traditional information security, the entry-level language for interacting with language models is natural language (at least for English), and so no coding skills are required. This has led to a huge effervescent and creative frontier where thousands of people attack language models in an attempt to make them misbehave. We investigate how people attack language models - what motivates them, what their techniques and strategies are, how to describe the activity, and how attacks are developed. And, of course, what the attacks are. New resources have emerged for automatically attacking language models and for collecting various attacks from the community, forming an active baseline for discovering language model vulnerabilities. We will discuss the issues involved in collating attacks on language models, and how to use this information to make model interaction safer.     🗓️ Date: 11.06.2024 🕒 Time: 16:30 - 17:30 📍 Location: Swiss Text Analytics Conference, Chur, FHGR Fachhochschule Graubünden     Don't miss this opportunity to gain valuable insights from one of the leading experts in the field. Join us to explore innovative ideas and engage in thought-provoking discussions. #Keynote #SwissTextAnalyticsConference #Event #Innovation #Networking #ThoughtLeadership #Learning #IndustryInsights Corsin Capol, Albert Weichselbraun, Mark Cieliebak, Manuela Hürlimann, Claudiu Musat, Elisabeth Maier

5

For all the wizardly power of Large Language Models (LLMs), they cannot reliably handle math. In our world of strategic insights, we need both deep conversations and rigorous statistics, integrated into one cohesive story. Tools like ChatGPT and Claude are game changers when it comes to analyzing the "conversation" part of the equation. To overcome their inherent weaknesses in reliably computing math and statistics, we utilize something they are now very good at, Python programming. By leveraging Python's robust libraries, LLMs can bypass their mathematical constraints and execute precise computations. Whether you're a data scientist, financial analyst, or researcher, you can learn how integrating LLMs with Python can enhance your analytical tasks, from simple calculations to complex forecasting models. Our latest piece shows how allowing LLMs to utilize Python programming democratizes access to advanced computational tools, enabling users with limited programming expertise to find insights from sophisticated mathematical and statistical techniques. #ArtificialIntelligence #MachineLearning #DataScience #PythonProgramming #LLMs #MathInAI #DataAnalysis https://lnkd.in/gq-QWqig

7

DE-CIX Frankfurt, one of the world’s largest Internet Exchanges, has set a new record for data throughput. At 9:21 p.m. (CEST) on April 16, during the high-tension quarter-final second leg of the UEFA Champions League, data transmission peaked at a staggering 17.09 terabits per second (Tbit/s). The 17 Tbit/s milestone translates to the simultaneous streaming of approximately 5.7 million high-definition videos. To put this in more tangible terms, if each bit of data were printed on an A4 page, the resulting stack would reach a height of over 420 kilometers – nearly 50 times taller than Mount Everest. This surge in data traffic represents a 20% increase since the start of 2023 and an almost 60% rise from early 2022. Over the past five years, data throughput at DE-CIX Frankfurt has nearly tripled, underscoring the explosive growth and critical importance of digital #connectivity in today’s world.

10

🚀 Exciting Announcement: 𝗦𝗼𝗹𝘃𝗜𝗧 is Now Open Source on Github! 🎉 Published in PNAS (Proceedings of the National Academy of Sciences) earlier this year, this innovative pipeline is now available to the community! You can read the full paper here: PNAS: Context-dependent design of induced-fit enzymes using deep learning. 🔬 𝗪𝗵𝗮𝘁 𝗶𝘀 𝗦𝗼𝗹𝘃𝗜𝗧? SolvIT is a cutting-edge computational tool designed to streamline protein solubility prediction and optimize enzyme design. By combining graph neural networks with powerful tools like AlphaFold2 and Rosetta, SolvIT helps researchers prioritize high-solubility candidates and create well-expressed enzymes efficiently. 💡 𝗞𝗲𝘆 𝗙𝗲𝗮𝘁𝘂𝗿𝗲𝘀: AI-Powered Solubility Predictions: Predict solubility with precision using our GNN-based model. Seamless Automation: Automate feature extraction, prediction, and result formatting with Snakemake. Open Source: Free for the scientific community to enhance and innovate further. 📂 𝗚𝗲𝘁 𝗦𝘁𝗮𝗿𝘁𝗲𝗱 Explore the code and documentation here: GitHub Repository Dive into the science behind it in our PNAS publication: DOI: 10.1073/pnas.2313809121 🔗 𝗪𝗵𝘆 𝗧𝗵𝗶𝘀 𝗠𝗮𝘁𝘁𝗲𝗿𝘀: Protein design is one of the most exciting frontiers in computational biology. Tools like SolvIT simplify the path to designing enzymes that are thermally stable, highly active, and expressible, unlocking new possibilities in industrial and therapeutic applications. This is just the beginning. We are excited to see how the community uses SolvIT to push the boundaries of protein engineering! Let’s solve it together with SolvIT. 💡✨ Link to the repository: https://lnkd.in/dA62-Dmk

53

2 Kommentare

At Frankfurt: Programming Related to Artificial Intelligence - Publishing Perspectives: At Frankfurt: Programming Related to Artificial Intelligence  Publishing Perspectives http://dlvr.it/TC141m #ai #artificialintelligence

1

AI Engineer – Rethink the system with AI and stay one step ahead now! In a new episode of #InnovationSnippets, Lara Fehr, Juliane Turzynski and Nianheng Wu talk about T-Systems' AI solution #AIEngineer, which programs and translates old programming languages into new ones - For example, 40-year-old FAME programs from the financial industry in current #Python code. It also automatically creates documentation and tests parts of the program. The AI Engineer has been used in daily work at T-Systems since 2023. If you want to unlock the full potential of your business with the simplicity and efficiency of AI, please contact us and book a workshop at [email protected]. #InnovationSnippets #innotastic #tsystems #innovation #AI #Metaverse

59

1 Kommentar

🔍 Don't Miss Out on This Insightful Session! 🔍 Join us at Data Week Leipzig for a dynamic presentation on ""Industrial Knowledge Graphs in action"" by Nikolay Krustev from eccenca's partner Ontotext 🕒 When: April 18th, between 13:00 - 14:30 📍 Where: Sitzungssaal 🎙️ Language: English Nikolay Krustev, leading Ontotext's Industrial Vertical, brings over a decade of experience in driving strategic business development across Automotive, Manufacturing, AECO, and Energy sectors. In this session, Nikolay will unveil real-world applications and use cases, demonstrating the power of knowledge graphs in transforming industries. From enhancing data product catalogs to architecting modern data stacks for commercial analytics, Nikolay's insights promise to inspire and enlighten. Don't miss this opportunity to delve into the world of industrial knowledge graphs and gain valuable insights from Nikolay and other industry experts. See you there! #KnowledgeGraphs #DataWeekLeipzig #IndustrialInsights #eccenca #Ontotext

23

𝗡𝗲𝘄 𝗗𝗘/𝗘𝗨 𝗢𝗽𝗲𝗻 𝗦𝗼𝘂𝗿𝗰𝗲 𝗟𝗟𝗠 🇪🇺 𝗧𝗲𝘂𝗸𝗲𝗻 𝟳𝗕, 𝗡𝗼𝘄 𝘁𝗿𝘂𝗹𝘆 𝗢𝗽𝗲𝗻 𝗦𝗼𝘂𝗿𝗰𝗲? 𝗦𝗵𝗼𝗿𝘁 𝗙𝗮𝗰𝘁𝘀 🇪🇺 European language focus 🤖 7B parameters 🔗 Very open Apache 2.0 license 🛠️ Instruction Fine-tuned 📊 Not outperforming in Benchmarks but also within normal distribution (for 🇬🇧) 👶🏼 4k max context 💣 (EDIT) Alignment Discussion see here: https://lnkd.in/eseXDMMS My former colleagues Mehdi Ali (roommate within our time in Smart Data Analytics at Rheinische Friedrich-Wilhelms-Universität Bonn of Jens Lehmann) and Stefan Wrobel (one of my Ph.D. supervisors Rheinische Friedrich-Wilhelms-Universität Bonn) and their teams have released an LLM developed in Europe within OpenGPT-X & Fraunhofer IAIS. This has been developed with a focus on European languages. I'm excited to try it out right away and see the actual performance! They announced their own Leaderboard and the Model in somewhere in between other models of same size: https://lnkd.in/eTgv8bjV 𝐌𝐲 𝐓𝐚𝐤𝐞 🤗 • I have high hopes for a better documentation of the development process and I am curious if this model can be called truly open source and not just open weight. • I look forward to further European and German initiatives that also pursue (Gen)AI development here in 🇪🇺❤️ 𝗠𝘆 𝗤𝘂𝗲𝘀𝘁𝗶𝗼𝗻𝘀 (at Mehdi Ali, Michael Fromm, Max Lübbering, PhD and all your contributors): • Is anyone capable to entirely reproduce the model from the documentation? • What do you expect from Arena Elo Scores? • In which tasks would you recommend usage of your model over other models from Mistral AI, Meta, Google, Microsoft, Alibaba Cloud like Mistral7B, LLama3.18B, Gemma, Phi, Qwen... • Will we see truly big LLMs in size of 50-400B Parameter that can compete with current top notch LLMs from your side? We at Comma Soft AG provide with Alan.de a European and German Solution to use GenAI in a highly performant, safe and trusted environment. If you like to see more of our recent R&D, check my recent posts and reach out to me for more idea exchange 😊

47

35 Kommentare

How to do fun stuff with LLMs: I gave a short introduction to Retrieval Augmented Generation at the University of Basel. Here are some of the slides. Hannes Grassegger had invited me to his media studies seminar. What I wanted to show the students: – How easy it is to talk to ChatGPT and a large text file with a Jupyter notebook and some Python. (Thanks Florian Leuerer for teaching me.) – How this works even easier with a no-code tool like VectorShift, and how to chat with a YouTube video. (Thanks to Nikita Roy.) – How to run a RAG locally, securely and for free on your own computer, with a program like AnythingLLM and different models. (Thanks again, Nikita.) I very much hope that this hands-on training demystified AI to some extent and showed an example of how this partly unreliable technology can be put to good use. Even for journalism. And that you can try out some of these things yourself pretty quickly. (I have already discovered that chunking and sorting results quickly becomes complicated again.)

59

1 Kommentar