René Salmon

🚀 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

What Karl says. There's going to be two amazing talks about Hopsworks based #ai and #ml systems: * Build a personalized Bitcoin (BTC) virtual assistant in Python with Hopsworks and LLM function calling by Javier de la Rúa Martínez (https://lnkd.in/eQfWpFRU ) * Build TikTok's Personalized Real-Time Recommendation System in Python with Hopsworks by Jim Dowling (https://lnkd.in/eXEHePjG) If you are in the #berlin area next week - go and take a look!

7

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

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

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 Kommentar

The Uncertain Art of Accelerating ML Models with Sylvain Gugger: Sylvain Gugger is a former math teacher who fell into machine learning via a MOOC and became an expert in the low-level performance details of neural networks. He’s now on the ML infrastructure team at Jane Street, where he helps traders speed up their models. In this episode, Sylvain and Ron go deep on learning rate schedules; the subtle performance bugs PyTorch lets you write; how to keep a hungry GPU well-fed; and lots more, including the foremost importance of reproducibility in training runs. They also discuss some of the unique challenges of doing ML in the world of trading, like the unusual size and shape of market data and the need to do inference at shockingly low latencies. You can find the transcript for this episode  on our website. Some links to topics that came up in the discussion: * “Practical Deep Learning for Coders,” a FastAI MOOC by Jeremy Howard, and the book, of which Sylvain is a co-author. * The Stanford DAWNBench competition that Sylvain participated in. * HuggingFace, and the Accelerate library that Sylvain wrote there. * Some of the languages/systems for expression ML models that were discussed: PyTorch, TensorFlow, Jax, Mojo, and Triton * CUDA graphs and streams * Hogwild concurrency #OCaml #OCamlPlanet

2

🚀 Exciting news from my colleague Andreas Wagner ! In his latest talk during Berlin Buzzwords, Andreas delves into the intricate world of measuring the effectiveness of search functionality changes. It's a persistent challenge, even for seasoned teams who often depend on search relevance feedback labels. While Web-scale Experimentation like A/B and Multivariate Testing is common, few companies leverage it to incrementally improve their search systems at the query level. A/B Testing in search is particularly complex, with no one-size-fits-all approach. It involves selecting the right evaluation metrics, determining suitable randomization units, and handling imbalanced and sparse data to make precise decisions—especially when data is limited. Over time, we've developed and refined our "Query Testing" capability. Andreas shares invaluable insights from this journey, shedding light on the nuances of search optimization. 🎥 Check out his talk here: bit.ly/4clQsbl

19

𝐔𝐧𝐥𝐞𝐚𝐬𝐡𝐢𝐧𝐠 𝐭𝐡𝐞 𝐏𝐨𝐰𝐞𝐫 𝐨𝐟 𝐆𝐞𝐧𝐀𝐈 𝐟𝐨𝐫 𝐄𝐧𝐞𝐫𝐠𝐲 𝐑𝐞𝐭𝐚𝐢𝐥 Was my Keynote topic at CIO Events Generative AI Summit in Munich this week. Here are my Top 3 takeaways: 1. 💶 𝐌𝐨𝐧𝐞𝐭𝐢𝐳𝐚𝐭𝐢𝐨𝐧 𝐨𝐟 𝐃𝐞𝐦𝐨𝐜𝐫𝐚𝐭𝐢𝐳𝐞𝐝 𝐆𝐞𝐧𝐀𝐈: Although more than half of the participants had in some for a <insert company name> GPT or similar platform, it remains difficult to monetize and fund aside from innovation budget. I'm therefore happy that we at E.ON have early on established a very successful license model for our internal "Chat with Your Data" (a LLM based RAG solution for specifc user groups) solution after democratizing GenAI by opening up our "Ask me Anything" (enhanced drop-in replacement for ChatGPT, but safe and secure) service to all 74k employees at E.ON. 2. 📈 𝐅𝐫𝐨𝐦 𝐇𝐲𝐩𝐞 𝐭𝐨 𝐇𝐢𝐠𝐡 𝐕𝐚𝐥𝐮𝐞 Now is the time to push high-value use cases forward, a company replacement for ChatGPT is only the start. NOW is the time to identify the highest value use cases. In Energy Retail the main cases I see revolve around the topic of customer service automation, three topics to highlight: automated email classification and information extraction, enabling customer service agents with natural language based RAG systems to get quick answers from their knowledge bases, and third automatic checks on bills and other documents. All of these are much easier now with LLM's. The marginal cost of asking LLMs to extract information is close to 1ct, this is 20-100 times cheaper than human labor on the same task! The APIs of our GenAI platfrom enable the cases! 3. 🤖 𝐖𝐡𝐞𝐫𝐞 𝐚𝐫𝐞 𝐚𝐥𝐥 𝐭𝐡𝐞 𝐂𝐨𝐏𝐢𝐥𝐨𝐭𝐬?  I'm not sure what your experience is, but the takeaway from the conversations for me was: many have tried especially the Office365 CoPilots, but hardly anyone is really convinced on their value including myself. The only CoPilot that was received positive by a majority were the GitHub CoPilots for development. What do you think about these? Special thanks to the organizers of the event, everyone for the great conversations and especially to Gerhard Holzwart for inviting me to speak! #data_on #welovetechnology #itsonus #ai #genai #generativai #cio

197

27 Kommentare

🔍 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

📊🚀"A Leap of Faith: Machine Learning Without Mastering Statistics" - a conversation between myself on D3M Labs and Dr. Ana Moya, Head of Data Science & Analytics at Handelsblatt Media Group & Lecturer at International School of Management, Germany. AI-FOMO, fear of missing out, is rampant in Germany. As companies are afraid of falling behind, they embark on AI projects without a firm grasp of statistical fundamentals or understanding their data. In this video we talk about how companies can avoid burning money on expensive one-off lighthouse projects and instead use data to add business value. We also talk about navigating build vs. buy discussions, healthy analytics project execution habits and creating a healthy data culture. How are you countering AI-FOMO to ensure long-term business value from data? Introduction: Why Companies Often Dive into AI Projects Hastily Without Mastering Statistical Foundation 2:31 - The danger of plunging into machine learning and neglecting the statistical principles. Tips for discussions with consultants and vendors. 3:42 - Why do many companies start directly with predictive analytics or even prescriptive analytics without understanding their data. How do you conduct different types of analysis? 6:16 - How can we get people excited about statistics? How do we make statistics and data literacy attractive? 7:42 - Secrets of success for AI and analytics projects: The importance of good habits. Follow D3M Labs to join the conversation. Subscribe to the D3M Labs YouTube Channel.

6

Whether you’re training an entire LLM oder try to stop your cat bringing mice home: Getting your AI model “just right” and streamlining the processes around it is 95% of the work. Only reality will reveal your model’s performance. You want robust data pipelines and training processes to get there as fast as possible. Iterate quickly to make the right corrections to the model early. Until the model is ‚just right‘. LMOps takes care of that, in fact of the whole lifecycle and together with the right platforms, is crucial for success in Machine Learning. Inspired by DevOps principles: Fast flow and early feedback. RAG (Retrieval Augmented Generation), which integrates your company data with large language models, is the newest kid in town and guess what: RAGOps is what makes it happen in a robust way. Something to watch out.

119

15 Kommentare

Andreesen Horowitz (a16z) has published a long piece on AI voice agents for B2B and B2C outlining challenges but also opportunities https://lnkd.in/eJXmwRZx. #aivoiceagents #aiagents #voicebot #generativeai #artificialintelligence #a16z

6

Wrapping up an amazing #PyConDE & #PyDataBerlin! Looks like our team really enjoyed having a discussion about #Streamlit with Lukas Masuch and Arnaud Miribel. Want to learn more? Missed out PyCon? No worries! We're hosting a meetup in Mainz next week to discuss key trends & takeaways. Stay tuned for details and our write-up. #Germany #DataScience #predict42, #NLP

25

1 Kommentar

An interview with me about the Zettelkasten methodology, in my app Conoted. How I implemented this methodology in my project and integrated Zettelkasten in ranking contacts for knowledge management

3

Exciting to see that folks are speaking our language! Thanks for the mention Neo4j 🚀 Check out our recent benchmark study cited in the blog post: https://lnkd.in/epJMkBQa

7

Semafor writes about Amazon AI research on how to get rid of bad AI data https://lnkd.in/edHVWWJR. #trainingdata #artificialintelligence #amazon #aws #semafor

6