New Geometry-Aware Operator Transformer (GAOT) combines speed and accuracy for PhysicsML applications. Link to Paper/code below 👇🧵

Learn why AI-accelerated, physics-based, high-fidelity surrogates for interactive digital twins are 65X cheaper and use 58X less energy than numerical simulations. #PhysicsML #EnergyEfficiency . Read More:

Just finished working on a independent blog paper on Physics Informed Machine learning on dynamical systems.Paper is complete.#PIML #PINN #NeuralODE #FNO #SINDy #ML #PhysicsML #DeepLearning.

Short part from my appearance at the Makers Bar podcast - discussing AI-driven simulation, #PhysicsML, building @siml_ai, and more. Full interview here:

Pretty excited about my upcoming talk about #physicsml #sciml at SlovakiaTech Expo/Forum — can’t wait to show folks some mindblowing use cases of AI-driven approaches to engineering and science, and the future where we’re heading!

We're working hard on adding an ability to build @holoviz Panel -based apps on top of @nvidia Modulus-based PhysicsML simulators. Doing CFD analysis on the mobile phone / tablet in the web browser with 2-second simulations of 10,000 variations of different inputs… while the

And also introducing PhysicsML - a package for physics-based/related models, providing a freely available, standardised and modular interface for handling, building and training 3d models with an emphasis on #neuralnetpotentials. Find out more:

Physics simulations? Yes, there's an AI for that. #PhysicsML #SciML

AI shouldn’t replace engineers, it should amplify their reasoning, insights & creativity. A must-read for anyone in AI + manufacturing. 📄 [arXiv:2506.00056].#AI #InverseDesign #PhysicsML #Manufacturing #LLM #DesignAutomation.

Today, I was surprised to finally get access to @OpenAI's Deep Research. Already tested it on multiple of tasks (preparation of grant proposal, public procurement proposal, competitors research, latest PhysicsML research review). Super impressed!

Slowly, one at a time… but within few years, I want to get @TheDimensionLab to a spot where most of these applications of PhysicsML could be done with the help of DimensonLab’s tooling or in some way connected to DimensionLab/Siml.ai.

再メモ.

Works on Quantum information + ML

Join us on April 26 at 8AM for a deep dive into NVIDIA's physics-ML framework, Modulus, as we discuss how #PhysicsML is changing what is possible. Register now.

Tensor Networks: Putting Quantum Wavefunctions into Machine Learning #AIUkraine

Looking for the latest research in scientific machine learning, PhysicsML, PINNs, FNOs, and similar topics?. Get it right here: We've added a new section called "Papers" to the @TheDimensionLab website, which lists the latest research papers from @arxiv

Papers | < physics | machine learning >.機械学習で物理なら色々ある

Check out #MachineLearning topological defects in the XY model, by Matt Beach. On the revamped <physics|machine learning> blog @_anna_go @MStoudenmire .