#PhysicsInformedML
LightNews LightNews
helioxpodcast.bsky.social
🕸️ Why AI's Next Breakthrough Isn't About More Data: #MKAN

www.buzzsprout.com/2405788/epis... 

helioxpodcast.substack.com/publish/post... 


#AI #MachineLearning #DataScience #ScientificComputing #TechInnovation #ComputationalScience #PredictiveModeling #DeepLearning #PhysicsInformedML
September 3, 2025 at 2:06 PM
jmlmc.bsky.social
Physics-Informed Encoder-Decoder Gated Recurrent Neural Network for Solving Time-Dependent PDEs

dl.begellhouse.com/journals/558...

#PhysicsInformedML #RecurrentNeuralNetwork #PDESolver
June 1, 2026 at 4:15 PM
chingyaolai.bsky.social
From #PhysicsInformedML to #MLInformedPhysics: We're excited about the "knowledge discovery" component of this project utilizing vast amount of Earth 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.💡
science.org Science Magazine @science.org · Mar 17
In Science, researchers report a physics-informed #DeepLearning model that can predict the deformation behavior of Antarctic ice shelves, revealing complexities of the process that extend beyond the traditional understanding.

Learn more in a new #SciencePerspective. scim.ag/3RgudLc
How does Antarctic ice deform?
A deep-learning model infers large-scale dynamics of Antarctic ice shelves
scim.ag
March 19, 2025 at 5:34 PM
jmlmc.bsky.social
Hybrid PDE-Deep Neural Network Model for Calcium Dynamics in Neurons

www.dl.begellhouse.com/journals/558...

#NeuralModeling #PhysicsInformedML #PDEs #DeepLearningInScience
May 13, 2025 at 3:13 PM
begellhouse.bsky.social
Hybrid PDE-Deep Neural Network Model for Calcium Dynamics in Neurons

www.dl.begellhouse.com/journals/558...

#NeuralModeling #PhysicsInformedML #PDEs #DeepLearningInScience
May 13, 2025 at 3:11 PM
ali-ameli.bsky.social
Check out our new paper in WRR: in the Prairie Pothole Region, millions of wetlands fill, spill, and connect, making runoff hard to predict. We show how physics + AI can improve prediction of streamflow and wetland storage in ungauged watersheds. #Wetlands #NewPaper #AGUPubs #PhysicsInformedML
Regionalization of Hydrologic Behavior and Pothole Water Storage Dynamics in the Prairie Pothole Region
Physics-informed deep learning outperforms LSTM in regionalizing hydrologic behavior and pothole storage across PPR (un)gauged catchments PPR-scale model captures inter(a)-annual variability of h...
agupubs.onlinelibrary.wiley.com
March 11, 2026 at 7:10 PM
jmlmc.bsky.social
64 citations: Multifidelity transfer learning trains deep CNNs using mostly cheap low-resolution simulations supplemented with select high-res runs—delivering expensive method accuracy at a fraction of the cost.

📖 www.dl.begellhouse.com/journals/558...

#PhysicsInformedML #DataScience
December 26, 2025 at 2:31 PM
scholarcast.bsky.social
New TurbRec: unsup. physics-consistent full-scale turbulence super-res. Low-res inputs + equations, no high-res data. Multi-scale Fourier embeddings; physics fixes small structures. Beats PINNs on spectra & stats.
link.springer.com/article/10.1...
#Turbulence #SuperRes #PhysicsInformedML
Physically-consistent full-scale unsupervised reconstruction in turbulent flow dynamics - Acta Mechanica Sinica
An advanced super-resolution reconstruction method for turbulent flows should be efficient, avoiding reliance on high-resolution data; reliable, strictly adhering to physical laws; and accurate, effec...
link.springer.com
September 14, 2026 at 11:07 AM
jmlmc.bsky.social
Role of Physics in Physics-Informed Machine Learning

dl.begellhouse.com/journals/558...

#PhysicsInformedML #DimensionalAnalysis #ScientificComputing
July 9, 2026 at 5:15 PM
getnews-me.bsky.social
A physics‑informed learning machine using an extreme learning network trains in about one second, enabling real‑time monitoring of soil‑pile deflection. https://getnews.me/physics-informed-learning-machine-improves-soil-pile-modeling/ #physicsinformedml #geotechnical
Physics‑Informed Learning Machine Improves Soil‑Pile Modeling
October 3, 2025 at 12:36 AM
getnews-me.bsky.social
Researchers introduced training that separates data‑driven and physics losses, using Traditional Multi‑Gradient Descent to improve traffic flow simulations. Read more: https://getnews.me/multi-gradient-descent-boosts-physics-informed-traffic-flow-models/ #physicsinformedml #trafficflow
Multi-Gradient Descent Boosts Physics-Informed Traffic Flow Models
September 20, 2025 at 9:51 AM
getnews-me.bsky.social
Researchers find that affine‑variety dimension—not model size—drives generalization in physics‑informed ML; linear and nonlinear PDE tests confirm the results. Read more: https://getnews.me/understanding-generalization-in-physics-informed-machine-learning-models/ #physicsinformedml #generalization
Understanding Generalization in Physics‑Informed Machine Learning Models
September 18, 2025 at 10:11 PM
begellengineering.bsky.social
Hybrid PDE-Deep Neural Network Model for Calcium Dynamics in Neurons

www.dl.begellhouse.com/journals/558...

#NeuralModeling #PhysicsInformedML #PDEs #DeepLearningInScience
May 13, 2025 at 3:12 PM
identityelement.bsky.social
Beyond metric preservation: build symplectic structural invariance into representation.

arxiv.org/abs/2512.19409

#ReservoirComputing #RepresentationLearning #InformationGeometry #SymplecticGeometry #HamiltonianDynamics #GeometricDeepLearning #DynamicalSystems #PhysicsInformedML
Symplectic Reservoir Representation of Legendre Dynamics
Modern learning systems act on internal representations of data, yet how these representations encode underlying physical or statistical structure is often left implicit. In physics, conservation laws...
arxiv.org
December 23, 2025 at 8:54 AM
informaq.bsky.social
Quantum circuits extract 42% more uncertainty information than classical ML methods with better calibration - promising safer, more reliable AI for medicine, autonomous vehicles, and critical systems.

#QuantumML #UncertaintyQuantification #PhysicsInformedML
Quantum Circuits Enable Efficient Uncertainty Quantification for Physics-Informed Machine Learning
iq.fp2.dev
April 29, 2026 at 12:46 PM
informaq.bsky.social
OAMNet, a FiLM-modulated CNN with soft OAM-conservation loss, reconstructs high-gain SPDC modal structure 128x faster than numerical simulation with >30% accuracy gains over U-Net baselines, enabling real-time quantum-optical characterization.

#QuantumPhotonics #PhysicsInformedML #Research
Physics-Guided Neural Networks for High-Dimensional Quantum Entanglement Characterization via SPDC
iq.fp2.dev
April 7, 2026 at 3:11 AM
jmlmc.bsky.social
How can AI improve vaccine distribution? New research uses physics-informed neural networks to estimate disease parameters & calculate optimal distribution for diverse populations—even with noisy data.

dl.begellhouse.com/journals/558...

#PhysicsInformedML #VaccineDistribution
January 26, 2026 at 3:18 PM
jmlmc.bsky.social
How to predict population dynamics with incomplete data? fPINNs provide a robust framework for fractional prey-predator models, accurately inferring predator populations from prey data alone under real-world constraints.

dl.begellhouse.com/journals/558...

#PhysicsInformedML #PopulationDynamics
January 20, 2026 at 2:19 PM
jmlmc.bsky.social
66 citations: Tensor basis GP models for hyperelastic materials build physical invariances into model structure, achieving better predictions with less data than black-box ML approaches.

📖 www.dl.begellhouse.com/journals/558...

#PhysicsInformedML #Materials
December 24, 2025 at 3:36 PM
arxivlens.bsky.social
Solving the Cosmological Vlasov-Poisson Equations with Physics-Informed Kolmogorov-Arnold Networks
Ashutosh Mishra, Emma Tolley et al.
Paper
Details
#Cosmology #PhysicsInformedML #KarnovArnoldNetworks
December 15, 2025 at 9:03 AM