HAMNO: New Neural Operator Architecture for Physics-Informed Learning of Complex Dynamical Systems
Researchers introduced HAMNO, a neural operator architecture designed to better learn solutions to partial differential equations involving multi-scale structures and long-range interactions. The model combines local convolutional and global spectral processing with an adaptive gating mechanism, and includes a physics-informed variant (PI-HAMNO) that incorporates PDE constraints. The work demonstrates improved accuracy and stability on benchmark equations compared to existing neural-operator baselines.
HAMNO addresses limitations in existing neural-operator architectures for modeling nonlinear time-dependent systems by integrating local convolutional representations, global spectral operators, and hierarchical encoder-decoder processing. A key innovation is a data-dependent gating mechanism that dynamically balances local and global information at each spatial location, enabling the model to capture fine-scale features while maintaining long-range dependencies. The physics-informed extension, PI-HAMNO, employs a multi-objective loss function combining data fitting with strong-form and weak-form physics constraints derived from the governing PDEs. Evaluation on non-periodic Allen-Cahn, Cahn-Hilliard, and Swift-Hohenberg equations demonstrates that HAMNO outperforms standard baselines across long-horizon predictions, limited-data scenarios, out-of-distribution initial conditions, and random-seed variations, with PI-HAMNO providing additional improvements in stability and data efficiency.
What's missing
The study does not discuss computational cost or scalability comparisons with baseline methods, nor does it address applicability to real-world experimental data or three-dimensional systems beyond the cubic domain test cases presented.
What different sources said
- arXiv cs.LGCenter
HAMNO: A Hierarchical Adaptive Multi-scale Neural Operator with Physics-Informed Learning for Dynamical Systems
Related
Genetic Drift, Not Selection, Drives Rapid Feather Color Evolution in Island Bird Radiation
A new study of an island bird radiation found that rapid evolution of feather coloration is driven primarily by genetic drift in small populations rather than sexual or ecological selection. The research integrated whole-genome data with detailed plumage measurements across complete species sampling to test whether signaling trait evolution correlates with speciation rates. The findings suggest that neutral demographic processes play a central role in generating phenotypic diversity during island radiations, challenging assumptions about the mechanisms driving rapid evolution.
New AI Model Improves Prediction of Therapeutic Peptide Function from Protein Sequences
Researchers developed a lightweight CNN classifier that predicts whether peptide sequences have therapeutic properties, trained on a database of 54,655 peptides across 48 functional categories. The model uses a novel negative sampling strategy to reduce false positive rates from over 60% in previous approaches to 2.1%. This advancement could accelerate drug discovery by enabling faster computational screening of peptide candidates before expensive experimental testing.
Study Shows Different Metabolic Stress Models Produce Distinct Effects on Human Neuronal Networks
Researchers tested three common in vitro metabolic stress models on human-derived neuronal networks and found each produced different patterns of neuronal activity and cell damage. The models tested were hypoxia alone, oxygen-glucose deprivation (OGD), and hypoxia combined with glutamate exposure. The findings suggest that choice of experimental model significantly affects results and that combining electrophysiological and structural analyses is important for accurately assessing metabolic stress in stroke research.