Finally published @ScienceMagazine: .Can AI help yield new insights from vast amounts of Earth data? .We use large-scale data and neural nets to find the constitutive laws of glacial ice, which differ from commonly assumed forms in conventional models.

We're pleased to release DIFFICE-jax v1.0, the foundation of our Science paper and the first #DIFFerentiable #NeuralNetwork solver for #DataAssimilation of ICE shelves written in #JAX: .🔗Docs: 📄Peer-reviewed by JOSS #OpenSource:

From #PhysicsInformedML to #MLInformedPhysics: We're excited about the "knowledge discovery" component of AI utilizing vast amount of data. There is much more out there to be discovered, as Nature's imagination is far greater than that of humans. Don't stop searching.💡.

RT @ScienceMagazine: In Science, researchers report a #physics-informed #DeepLearning model that can predict the deformation behavior of An….

NASA Earth data + AI enable us to infer the constitutive models critical for ice dynamics. AI is useful, but no data = no discovery. Below is an example of the training data: a velocity map showing the dynamics of the Antarctic Ice Sheet. Source: #NASA.

Looking forward to learning about recent advances in #AI4Climate at the @APSphysics #GlobalPhysicsSummit. Come check out the back-to-back focus sessions, "AI Applications in Weather and Climate I & II," on Tuesday from 9:00 AM to 1:30 PM!.

RT @AlbanSauret: Excited to share our new paper in @PNASNews Ice cubes often appear cloudy because, as water freez….

This project would not have been possible without postdoc Yongji Wang, who has done a tremendous job leading this work over the past four years, making the impossible possible. 💪🧊.#ScienceResearch @stanforddoerr.

Our open-source JAX package, DIFFICE.jax (DIFFerentiable neural-network solver data assimilation of ICE shelves) + a 60 page SI, are released. We are actively expanding the code's applications to other datasets and welcome collaborations!.

Special thanks to Bryan Riel for writing a beautiful perspective on our paper, summarizing the key nuances of both our scientific findings and algorithmic advances.

Result: We find the constitutive law of glacial ice near grounding zones follows power laws but with varying exponents, which can impact grounding line stability. We also develop a method to detect and infer ice's anisotropic viscosity—a long-hypothesized property of glacial ice.

Method: We inject only the physics that we believe are correct, e.g. conservation of momentum, into the ML training, and leaving the uncertain physics, e.g. the constitutive law, to be determined through optimization against observations. Details documented in our 60-page SI.

RT @ZLabe: After nearly two weeks of overwhelming uncertainty, today it happened. I was fired from my dream of working at NOAA. I'm so sorr….

RT @ParisPerdikaris: 🎉Excited to announce major new breakthroughs in our Aurora foundation model! Our team (@crisbodnar, @ikwess, @a_lucic,….

Our new review paper "Machine Learning for Climate Physics and Simulations" with Pedram @turbulentjet, Aditi, Maike, Raffaele & Balaji is online @AnnualReviews 🎉. Beyong accelerating simulations, can ML help us understand climate physics? We highlight the recent progresses &