Farshad Saberi Movahed, PhD’s Post

Are you joining the NVIDIA GTC Conference 17 - 21 March? The #1 AI event for developers and all of us who are AI curious! It´s online! Dell Technologies will be there? Session on 20 March: Turbocharging NVIDIA DGX SuperPOD With Dell Data Lakehouse and PowerScale presented by Jay Limbasiya, Dell Technologies. Jay will talk about: Your AI is only as good as your data. Discovering, accessing, preparing, and serving data for AI remains complex. Organizations are struggling with data silos, proprietary formats, lack of secure access, and inefficient storage architecture to properly "feed the beast" that is your NVIDIA DGX SuperPOD. Join us to learn how Dell is solving these common “extract, transform, load (ETL)” problems by bringing an open, modern data lakehouse and a trusted NVIDIA DGX SuperPOD built on Ethernet technology and Dell PowerScale AI-Ready data platform. Martin Westblom Rasmus Bisgaard Ralph Stocker Arild Hansen Odd-Rune Moltu Kim Haurum Poul Kjeldgaard Jay Limbasiya Vibeke Østensvik Jonas Dietz-Olsen Christian Lai Finn Henningsen Roger Ihlebekk Kaj Inge Skjonhaug David Westerik Use the link below for free of charge online sign up.

Here is a session you won't want to miss at NVIDIA's GTC this month. Dr. Wei Wei from Accenture describes building compact large language models to handle specific tasks. If you haven't registered for GTC yet, click the link this link to register. https://lnkd.in/g-asYK7c The Art & Science of Evaluating, Customizing, and Building ML Models [S61754] Wei Wei, Denise Zheng, Lan Guan, Karthik Narain, Tom Stuermer, Azita Martin, Jay Puri, Shanker Trivedi, Alvin Da Costa, Marc Hamilton, Diane Genova, Shelley Wanner Guthmiller, Kim Ng, Osama Qazi, Sumanta Paul, Chantal Garcia Fischer, Mili Iyengar, Deepak Chaudhary, Teemu Nordman, Tunc Yorulmaz, Forrest Bussler, Viveca Pavon-Harr, Suzanne Randall, Trevor Brennan, Brad Zalischuk, #gtc24, Michael Shields

In the first 30 minutes of Jensen Huang's keynote at GTC, Jensen outlined the current fundamental concepts and innovations of high performance compute for AI. Must watch for any technologist and anyone who is interested to understand what is happening in AI. Here comes Blackwell GPU and the Superchip with two Blackwell GPUs +dies+ one Grace CPU to achieve complete coherence. Then the racks with Superchips to make the Blackwell system platform. https://lnkd.in/eErJQH7m Key themes: 📌 Superchip, compute platform, and data center architecture built around the next-gen chips after Hopper 📌 Parallelization in AI Computer 📌 Cloud platform and application agnostic 📌 Nvidia AI compute x Industries (manufacturing, climate, biology) 📌 Use-case and task-specific AI inference microservices (effectively customized generative models packaged and optimized on Nvidia AI software and compute stack CUDA) 📌 Multimodal capabilities to push robotics to the next level Nvidia announcement will likely make many current startups focused on native GenAI struggle as it is clear that Nvidia will look to make its AI compute and services more accessible to further capture market share. My posting reflects my opinion and is not investment advice #AI #nvidia #compute #largelanguagemodels #genai #llm #innovation #highcompute #bigdata #multimodalai

💥 One week out from #nvidiagtc ! I've rarely seen so many old friends and colleagues from so many different industries attending a single conference. I'll be there all four days of the conference so I decided to check out some of the talks ahead of time to find my top 5. Featuring speakers from Berkeley Lab, Tomorrow.io, U.S. Department of State, Emerson Collective, and of course NVIDIA this is what I came up with: Huge Ensembles of Weather Extremes using NVIDIA's Fourier Forecasting Neural Network (FourCastNet) https://lnkd.in/gaKPij7q Large-Scale Graph GNN Training Accelerated With cuGraph https://lnkd.in/gS6j3AVQ Bridging the Compute Divide to Mitigate Climate Risk https://lnkd.in/g6KZEXpM Heat Transfer Modeling for EV Batteries https://lnkd.in/gxRFfT8Q Global Strategies: Startups, Venture Capital, and Climate Change Solutions https://lnkd.in/geQdDAiS

🚀 Don't miss this session at NVIDIA GTC: "A Hacker's Guide to using GenAI and Software-Defined Radio for RF Spectrum Exploration" [S62508] 🛠️ Join Dustin Van Stee and Joshua Martinez as they unveil how to harness a $40 software-defined radio with NVIDIA's cutting-edge GenAI tech for RF signal exploration. Learn how to capture & analyze RF signals on a consumer grade SDR, streamline processes with GPU-acceleration, and translate audio to text for deep analysis using NVIDIA's Holoscan, Riva, and NeMo. 🌐 This talk is aimed at innovators in the Public Sector and is a treasure trove for tech enthusiasts eager to explore the future of RF spectrum analysis and AI. 📅 Don’t miss out on diving deeper into the synergy of SDR and GenAI. Register now: https://bit.ly/49v10n3 #NVIDIAGTC #SDR #GenAI #RFExploration #Innovation #PublicSector

I just saw this great NVIDIA GTC talk: Vision AI demystified [S62607]. You might know most of that already as well but might good to get a more technical overview over most of the current model architectures. https://lnkd.in/ehGFhMTh #GTC2024

Last call for another great GTC starting this week! This edition will be as usual packed with hundreds of sessions and here is my personal pick. If not done already, don’t forget to register for free now to get access to replays after the event! Full session catalog and registration below : https://lnkd.in/ejVHst4x NVIDIA 2024 GTC KEYNOTE – March 18 - 21:00 CET   ·        Solving the Generative AI Infrastructure Challenge in 2024 [S62227] ·        Digitalizing the World's Largest Industries with OpenUSD and Generative AI [S62783] ·        Simulating factories of the future with Hyper-Scale, Hyper-Reality Digital Twins ·        Transforming Discrete Manufacturing with Hexagon Smart Digital Reality and NVIDIA Omniverse [S62623] ·        Omniverse-Based Fab Digital Twin Platform for Semiconductor Manufacturing [S62610] ·        Generative AI Powered with Extreme-Scale Computing to Discover Novel Metal-Organic Frameworks for Carbon Capture at Scale [S62172] ·        Accelerating Simulations of Multiscale Chemical Reactors using NVIDIA Modulus [S62060] ·        Using Next-Generation Geospatial Applications and 3D Visualizations to Improve Climate Risk Resilience Strategies [S62188] ·        Unlocking the Power of Large Language Models: Challenges and Considerations of Retrieval-Augmented Generation in the Power Sector [S63301] ·        Enabling More Resilient Grids With Physics and Data-Driven ML Virtual Sensors [S62524] ·        GPU-Accelerating Process Simulation Performance using NVIDIA’s cuDSS Sparse Linear Systems Solver [S62515] ·        The Increasing Complexity of the Electric Grid Demands Edge Computing [S61864] ·        Applying an Electric Power Cognitive Large Model in Power Equipment Operation [S62278] ·        A Smart Robot Born in the Industrial Metaverse, Enabled by OpenUSD [S62687] ·        Omniverse Impactful Scientific Visualization using Art and Science [S62423]

#GenerativeAI promises to revolutionize every industry it touches - all that’s been needed is the technology to meet the challenge. NVIDIA founder and CEO Jensen Huang on Monday introduced that technology - the company’s new #Blackwell computing platform - as he outlined the major advances that increased computing power can deliver for everything from software to services, robotics to medical technology and more. You can see full #GTC2024 Keynote featuring 𝐉𝐞𝐧𝐬𝐞𝐧 𝐇𝐮𝐚𝐧𝐠, 𝐅𝐨𝐮𝐧𝐝𝐞𝐫 & 𝐂𝐄𝐎 of NVIDIA here: https://lnkd.in/ekdnHmHh 𝐀𝐧𝐝 𝐝𝐨𝐧'𝐭 𝐟𝐨𝐫𝐠𝐞𝐭 𝐭𝐡𝐞 𝐦𝐞𝐞𝐭𝐮𝐩 𝐰𝐢𝐭𝐡 𝐆𝐫𝐞𝐞𝐧𝐍𝐨𝐝𝐞! 🤖 Topic: What's happening in SE Asia? From H100 GPU access to LLMs or different languages to SD, chatbot 🗓 Date: March 19th 🕐 Time: 1 PM PDT 📍 Location: SJCC Level 1 🔗 RVSP: https://lnkd.in/g7rTYBcT See you there at #NVIDIAGTC! #GreenNode #GTC2024 #NVIDIA #Keynote #AI #Cloud #SustainableComputing

Excited to share this upcoming session at NVIDIA's GTC conference where my friend and coworker Patty Delafuente from NVIDIA will be speaking alongside colleagues from the IRS: "From Paper to Pixels: The TAO Toolkit Approach to Digitalization" https://bit.ly/3Pc5t6t In this session, they'll demonstrate an innovative AI workflow using NVIDIA RAPIDS and the TAO Toolkit to automate processing of handwritten forms and applications. This approach has the potential to dramatically improve efficiency and accuracy compared to manual data entry. Having worked in OCR at USPS previously, I know firsthand the challenges of dealing with diversity in handwriting styles. The TAO Toolkit seems to offer a promising solution by leveraging cutting-edge AI. The speakers will showcase how this was initially applied to help the IRS explore alternatives to manually processing millions of paper forms each year - a situation that often led to operational inefficiencies and improper payments. If you're interested in AI for document digitization and automation in the public sector, this is definitely a session to check out at GTC! Let me know if you'll be attending as well.

Just stepped out of NVIDIA’s March 2024 keynote, and wow, I’m still reeling! 🚀 NVIDIA unveiled the Blackwell architecture, promising monumental leaps in AI and computing power. Offering up to 2.5x the performance in training and 5x in inference compared to its predecessor. NVIDIA’s foray into generative AI is shaping a future where creativity meets computational power, making the impossible seem possible. And the NVIDIA GB200 Grace Blackwell Superchip? It’s the heart of the beast, powering the DGX SuperPOD to new heights of AI supercomputing power. The innovation on display wasn’t just impressive; it was visionary, signaling a new era where NVIDIA isn’t just participating in the future—they’re actively shaping it. Can’t wait to see how these breakthroughs transform our digital and physical worlds! #NVIDIA #AIRevolution #FutureIsNow