Altair’s Post

Austin Dulaney, Head of Machine Learning & Data Science at Liminal, will be joining us at Battery Forum to lead a discussion on Learning to Listen: Pitfalls to Avoid & Opportunities to Thrive Using ML in Battery Metrology (Ultrasonic)! RSVP HERE: https://lnkd.in/erycNxkK Dr. Austin Dulaney is the Head of Machine Learning and Data Science at Liminal Insights, an ultrasound-based battery manufacturing inspection and intelligence company. Austin has a background in creating physics-informed and ML models to study and understand complex materials and their behaviors. There is a pressing need for novel inspection methods and advanced data utilization strategies as battery demand and production grow rapidly. This talk will focus on some of the challenges and learnings around introducing machine learning-based inspection solutions into battery manufacturing. #batteries #metrology #manufacturing #machinelearning

🎉 Exciting Update! Introducing OpenMM 8 with pioneering machine learning integration. Spearheaded by our CEO Gianni De Fabritiis and Senior Scientist Raimondas Galvelis, this release enables enhanced precision in molecular simulations. Read more about it here https://lnkd.in/e2b667Zj #OpenMM8 #Acellera #MachineLearning

Our paper with Alex Johnston on using a photonic system (e.g. a lattice of exciton-polariton condensates) as an all-optical platform for the generative diffusion models of machine learning has just been published by the Physical Review Letters https://lnkd.in/ebYY_D7t

While it’s great to see the valuation of the company continue to increase dramatically, it’s even better knowing that this is because our solution continues to solve major problems and complexities that the industry has been struggling with for decades. 🚀 📈 👩🚀 #vastdata #sales #data #datastorage

Accelerating material characterization: Machine learning meets X-ray absorption spectroscopy https://buff.ly/4bip0em

😟 The challenge: An automotive company running a segmentation model struggled to achieve the target latency in production. 😀 The solution: Using Deci’s AutoNAC engine, powered by neural architecture search, a faster and smaller model was generated. Latency was reduced by 2.1x, model size was reduced by 3x, and memory footprint was reduced by 3x – all while maintaining the original accuracy. Have a computer vision use case that you want to optimize? Learn more > https://lnkd.in/gn893s5 #deeplearning #computervision #semanticsegmentation

Data-driven models have great potential to characterize complex systems. However, physics-based models are extremely important as they are able to summarize tons of information from experiments, solid theory, experts, and common sense. Also, physics-based models typically come with easily interpretable parameters that have physical meaning. Why throw that knowledge away when training data-driven models? our proposed augmented physics-based models (APBMs) leverage both worlds. Our recent article trades off both modeling options and presents promising solutions in between, whereby physics-based models are augmented by data-driven components. Check out "Augmented physics-based machine learning for navigation and tracking" in the Transactions of the IEEE Aerospace & Electronic Systems Society, coauthored with my esteemed colleagues Tales Imbiriba, Ondrej Straka, and Jindrich Dunik https://lnkd.in/eFhE6EkP #datadriven #physicsinformed #tracking #navigation

For my weather geek friends!! Discover how NVIDIA's FourCastNet algorithm leverages #ML to replace traditional numerical simulations for studying extreme weather events, offering accurate and efficient results at a lower computational cost. #GTC24 Register now to hear more from Berkeley Lab's William Collins.