AI topics’ Post

A key novelty lies in using Graph Neural Networks, which allow points to "communicate" and self-optimize for better uniformity. This approach marks a pivotal enhancement for simulations in robotics, finance, and computational science. https://lnkd.in/gKEaiMn9

Dear All, This is a gentle reminder about tomorrow's talk by Prof. Marta D'Elia from Stanford University, USA, as part of the Param-Intelligence (π) Seminar Series. The seminar will take place on December 5, 2024, from 12:00 to 1:00 PM ET. Please mark your calendars and join us for this session. The Zoom link is provided below. Details of the talk are as follows: Title: On the use of Graph and Point networks in scientific applications Abstract: In the context of scientific and industrial applications, one often has to deal with unstructured space-time data obtained from numerical simulations. The data can be either in the form of a mesh or a point cloud. In this context, graph neural networks (GNNs) have proved to be effective tools to reproduce the behavior of simulated data; however, depending on the physical nature of the datasets, variations of vanilla GNNs have to be considered to ensure accurate results. Furthermore, when only a point cloud is available, one can also consider a graph-free approach by building a "point network" that doesn't require connectivity information. In this presentation we focus on particle-accelerator simulations; a computationally demanding class of problems for which rapid design and real-time control are challenging. We propose a machine learning-based surrogate model that leverages both graph and point networks to predict particle-accelerator behavior across different machine settings. Our model is trained on high-fidelity simulations of electron beam acceleration, capturing complex, nonlinear interactions among macroparticles distributed across several initial state dimensions and machine parameters. Our initial results show the model’s capacity for accurate, one-shot tracking of electron beams at downstream observation points, outperforming baseline graph convolutional networks. This framework accommodates key symmetries inherent in particle distributions, enhancing stability and interpretability. Speaker's Biography: Marta D'Elia is the Director of AI and ModSim at Atomic Machines and an Adjunct Professor at Stanford University, ICME. She previously worked at Pasteur Labs, Meta, and Sandia National Laboratories as a Principal Scientist and Tech Lead. She holds a PhD in Applied Mathematics and master's and bachelor's degrees in Mathematical Engineering. Her work deals with design and analysis of machine-learning models and optimal design and control for complex industrial applications. She is an expert in nonlocal modeling and simulation, optimal control, and scientific machine learning. She is an Associate Editor of SIAM and Nature journals, a member of the SIAM industry committee, the Vice Chair of the SIAM Northern California section, and a member of the NVIDIA advisory board for scientific machine learning. #SciML #DeepLearning Zoom link is available here:

This article explores groundbreaking research by the University of Michigan on mechanical neural networks (MNNs), materials capable of learning and performing computations. By adapting the backpropagation algorithm used in digital neural networks, researchers trained physical lattices to respond to inputs like force and modify their properties to achieve desired outputs. These 3D-printed lattices, modeled after neural connections, adjust segment stiffness to solve tasks such as species identification or complex mechanical responses. This innovation hints at a future where materials, like airplane wings, could autonomously optimize themselves for changing conditions. The approach also offers insights into biological learning processes and paves the way for more advanced applications using materials like polymers and nanoparticle assemblies. For more details, please continue reading the full article under the following link: https://lnkd.in/eDCMh33g -------------------------------------------------------- In general, if you enjoy reading this kind of scientific news articles, I would also be keen to connect with fellow researchers based on common research interests in materials science, including the possibility to discuss about any potential interest in the Materials Square cloud-based online platform ( www.matsq.com ), designed for streamlining the execution of materials and molecular atomistic simulations! Best regards, Dr. Gabriele Mogni Technical Consultant and EU Representative Virtual Lab Inc., the parent company of the Materials Square platform Website: https://lnkd.in/eMezw8tQ Email: gabriele@simulation.re.kr #materials #materialsscience #materialsengineering #computationalchemistry #modelling #chemistry #researchanddevelopment #research #MaterialsSquare #ComputationalChemistry #Tutorial #DFT #simulationsoftware #simulation

This article explores groundbreaking research by the University of Michigan on mechanical neural networks (MNNs), materials capable of learning and performing computations. By adapting the backpropagation algorithm used in digital neural networks, researchers trained physical lattices to respond to inputs like force and modify their properties to achieve desired outputs. These 3D-printed lattices, modeled after neural connections, adjust segment stiffness to solve tasks such as species identification or complex mechanical responses. This innovation hints at a future where materials, like airplane wings, could autonomously optimize themselves for changing conditions. The approach also offers insights into biological learning processes and paves the way for more advanced applications using materials like polymers and nanoparticle assemblies. For more details, please continue reading the full article under the following link: https://lnkd.in/eDCMh33g -------------------------------------------------------- In general, if you enjoy reading this kind of scientific news articles, I would also be keen to connect with fellow researchers based on common research interests in materials science, including the possibility to discuss about any potential interest in the Materials Square cloud-based online platform ( www.matsq.com ), designed for streamlining the execution of materials and molecular atomistic simulations! Best regards, Dr. Gabriele Mogni Technical Consultant and EU Representative Virtual Lab Inc., the parent company of the Materials Square platform Website: https://lnkd.in/eMezw8tQ Email: gabriele@simulation.re.kr #materials #materialsscience #materialsengineering #computationalchemistry #modelling #chemistry #researchanddevelopment #research #MaterialsSquare #ComputationalChemistry #Tutorial #DFT #simulationsoftware #simulation

✨ Excited to share our latest publication in Computer Methods in Applied Mechanics and Engineering! Audrey Olivier, Nicholas Casaprima, and I have developed a novel approach for incorporating functional priors into Bayesian neural networks via anchored ensembling, with applications in surrogate modeling for mechanics and materials. It bridges the gap between parameter-space and function-space priors, offering enhanced accuracy and uncertainty quantification in both in-distribution and out-of-distribution data scenarios. I am deeply grateful to my advisor, Dr. Audrey Olivier, for her invaluable guidance, encouragement, and expertise throughout this research. Key points: ✅ Traditional NNs are powerful but often deterministic and lack robust uncertainty quantification. Our approximate Bayesian approach enables better handling of epistemic uncertainties due to limited training data. ✅ By introducing a novel training scheme, our method leverages low-rank correlations in NN parameters to efficiently transfer knowledge from function-space priors to parameter-space priors in Deep Ensembling. ✅ We demonstrate that accounting for correlations between NN weights, often overlooked, significantly enhances the transfer of information and the quality of uncertainty estimation. 📊 Results: - A 1D example illustrates how our method excels in both interpolation and extrapolation scenarios. - A multi-input–output materials modeling case highlights its superior accuracy and uncertainty quantification for both in-distribution and out-of-distribution data. 📄 Read the paper: https://lnkd.in/gVKwvTnj 💻 Code and data: https://lnkd.in/gKQgDY-8 #BayesianNeuralNetworks #DeepLearning #FunctionalPriors #UncertaintyQuantification #SurrogateModeling

HEMANTH LINGAMGUNTA Revolutionizing Antimatter Production with Advanced Technologies Exciting breakthroughs in antimatter production are on the horizon, leveraging the latest advancements in quantum computing, machine learning, and particle physics. By combining quantum algorithms, neural networks, and deep learning techniques, we can optimize particle accelerator designs and extraction processes to dramatically increase antimatter yields[1][3]. Key innovations include: • Quantum-enhanced particle simulations for accelerator optimization • AI-driven beam control and focusing systems • Neural interfaces for precise accelerator tuning • Blockchain-secured data management for experiment integrity • Cloud-based distributed computing for complex calculations • Robotic systems for safe handling of antimatter particles • Advanced cybersecurity protocols to protect sensitive research These technologies could enable the production of usable quantities of antimatter for revolutionary applications in space propulsion, medical imaging, and fundamental physics research[5][7]. The future of antimatter science is bright. Join the conversation on how we can harness these cutting-edge tools to unlock the potential of antimatter! #AntimatterTech #QuantumComputing #AIforScience #SpacePropulsion #ParticlePhysics Citations: [1] Antimatter Quantum Interferometry - MDPI https://lnkd.in/gBEZfB9P [2] [PDF] Antimatter Production at a Potential Boundary https://lnkd.in/gzQrukDD [3] Quantum Computing and Simulations for Energy Applications https://lnkd.in/gzmBBcNU [4] MiniCERNBot Educational Platform: Antimatter Factory Mock-up ... https://lnkd.in/gx4Pwr7w [5] Steven Howe Breakthroughs for Antimatter Production and Storage https://lnkd.in/gyMvH5tF [6] Using AI to unlock the secrets of antimatter | by Ari Joury, PhD https://lnkd.in/g_VrdtMB [7] AEgIS Experiment Breakthrough: Laser Cooling Opens Door to New ... https://lnkd.in/g42Kjs8b [8] Artificial Intelligence in the world's largest particle detector https://lnkd.in/gBwdxb2G

🚀 KAIST Researchers Use #AI to Accelerate Quantum Simulations Exciting advancements are coming from #KAIST! The Quantum Insider shared that KAIST Researchers have developed an AI-based method that significantly reduces computation time for complex quantum simulations, making large-scale simulations faster and more accessible. 🔹 Breakthrough Highlights ✅ Streamlined Calculations: This new method bypasses traditional iterative processes commonly required in #DensityFunctionalTheory (#DFT) calculations. ✅ DeepSCF Model: Using a 3D convolutional neural network, the DeepSCF model learns chemical bonding information, effectively replacing the self-consistent field (#SCF) process. ✅ Enhanced Efficiency: By reducing computational load, this method offers faster quantum mechanical calculations, enabling researchers to tackle more complex, large-scale simulations. This is a significant step forward in making quantum simulations more efficient and accessible for industries and research fields that rely on advanced quantum mechanics. 🔗 Read the Full Article: https://lnkd.in/eUzuXUcu 💬 Thoughts on AI in Quantum Mechanics? How do you see AI transforming the field of quantum research? Follow for more such updates on quantum and AI advancements, and share this post to spread the word about this impactful development from KAIST! #QuantumMechanics #AI #QuantumComputing #QuantumSimulations #DFT #DeepLearning #Research #Innovation #SCF #KoreaAdvancedInstitute

Washington D.C - 05/16/2024 Title : [MIT Researchers Revolutionize Phase Transition Classification with Cutting-Edge AI Technology]. Researchers at MIT have pioneered an advanced generative AI-driven methodology to classify phase transitions in materials and physical systems with unprecedented efficiency. This physics-informed technique significantly surpasses the performance of traditional machine learning approaches, marking a pivotal advancement in the field. https://lnkd.in/eQWW4KAQ

[WSY](Rego)2021(Araújo)《Nonlinear Control System with Reinforcement Learning and Neural Networks Based Lyapunov Functions》(CLF)◆★ 【IEEE LATIN AMERICA TRANSACTIONS, VOL. 19, NO. 8, AUGUST 2021】‧可TRACE自(無PDF)〔https://lnkd.in/gs5XvX-x 〕【〔rosana.rego@ufrn.edu.br〕●Rosana C.B. Rego〔Process Control Laboratory, Federal University of Rio Grande do Norte, Natal, RN, 59077080, Brazil〕 】【〔meneghet@dca.ufrn.br〕●Fábio Meneghetti U. de Araújo 〔Department of Computer Engineering and Automation, Federal University of Rio Grande do Norte, Rio Grande do Norte, Brazil〕】【■Abstract≒This paper deals with the problem of finding the control Lyapunov function that keeps the system stable. 】【To find the Lyapunov function, this paper proposes the use of reinforcement learning with two neural networks based on the Lyapunov stability theory. 】【The proposed control is applied in two nonlinear systems. 】【The simulations show the good performance of the proposed technique and proved that reinforcement learning and neural networks are an excellent mathematical tool to deal with control design problems.】

This video highlights my ongoing MASc Thesis at U of T in the department of ECE supervised by Prof. Stephen Brown and Prof. Kevin Truong. Big thank you to Aman B. a PhD candidate at Caltech CNS in the Thomson group, for editing this video, inspiring many initial ideas for the project, and for helping through several breakthroughs this past week! Aman is a fellow EngSci grad from U of T! Also thank you to the Society for AGI for some very helpful discussions! A big inspiration for this project was the paper “Growing Neural Cellular Automata” (https://lnkd.in/gVcfejvn). The authors put a twist on Von Neumann cellular automata by demonstrating the interaction rules could be learned using neural networks to self-organize into desired 2D pixel patterns. Stephen Wolfram had already shown very simple cellular automata to be Turing Complete, so it was only a matter of time before their full possibilities started being explored. Moving from 2D to 3D, and fixed grids to free-moving cells required the insights of geometric deep learning, beautifully outlined in this proto-book: https://lnkd.in/gV4Yybk7. Geometric deep learning revolves around the idea of symmetry—which I think of as “the ways that something can be transformed without changing.” For example, a circle remains exactly the same when rotated in place. In a very real sense, that's what makes it a circle. In the more familiar case of bilateral symmetry, it's reflections over an axis that leave the object unchanged. Symmetry is the concept behind revolutions in modern mathematics (eg. group theory), physics (eg. gauge theory), and now deep learning (in architectures exploiting invariance and equivariance). CNNs led to Group CNNs (https://lnkd.in/g7AkqSK2), and eventually to the SE(3) Transformer (https://lnkd.in/gcV7MVzc). SE(3) is the special euclidean group, consisting of rigid translations and proper rotations in 3 dimensions—meaning this group inherently captures the structure of euclidean space. An architecture with SE(3) equivariance is indispensable for moving from fixed grids to freely moving cells. The power of Equivariant transformers can be seen through their use in AlphaFold, which outputs the 3D coordinates of all heavy atoms for a given protein (https://lnkd.in/gMSPc8t4...) AlphaFold recently won the Nobel Prize in Chemistry for its groundbreaking advance in protein structure prediction. The video of the frog developing was taken by Jan Van Ijken (https://janvanijken.com). Other inspirations include: Michael Elowitz, Rob Phillips, Karl Friston, John Conway, Carver Mead, Geoffrey Hinton, Richard Feynman, Eric Winfree, John Hopfield. I briefly foreshadowed a little bit of this project on the MLST Podcast a few months back (second half of that episode), and you can find that interview here (https://lnkd.in/gvXhKMKk)