SLE-FNO achieves zero forgetting and strong plasticity-stability balance in continual learning for FNO surrogate models of pulsatile blood flow by adding minimal single-layer extensions across four out-of-distribution tasks.
arXiv preprint arXiv:2505.22904 (2025)
11 Pith papers cite this work. Polarity classification is still indexing.
11
Pith papers citing it
citation-role summary
background 2
citation-polarity summary
roles
background 2polarities
background 2representative citing papers
A canonical engineering graph representation combined with region-aware graph attention learning enables robust and transferable 3D mode shape classification across heterogeneous vehicle models and sensor layouts.
A hybrid transformer-FEM integrator provides provable discrete energy preservation and gradient bounds for stable autoregressive forecasting of chaotic systems, with 65x fewer parameters and 9000x speedup in a fusion surrogate trained on 12 simulations.
Vision transformers trained on a new global dataset of Landsat-Sentinel-2 patches detect floating coastal algal blooms with 8-65% omission/commission error and outperform spectral indices under cloud and glint conditions.
A conceptual argument that data-driven scientific models generalize outside their training data only when their mathematical form matches the true governing equation, which rules out neural operators and neural boundary-value models.
A component-based reduced-order modeling framework decomposes multi-injector rocket combustors into trainable sub-models that couple to predict combustion dynamics across flow and geometry changes.
A multi-agent LLM framework, ATHENA, autonomously designs and refines numerical PDE solvers and physics-informed models, reportedly beating expert-authored baselines.
Reviews linear and nonlinear SciML surrogates for coupled fluid flow and transport, with new PINN modeling of turbidity currents and β-VAE mode extraction from Rayleigh-Bénard convection.
A replay-based continual learning strategy for physics-informed neural operators mitigates catastrophic forgetting on prior physical problems while enabling efficient adaptation to new data using only physical constraints.
A graph learning framework turns heterogeneous 3D engineering data into physics-aware graphs processed by GNNs for CAE mode classification and CFD field prediction in automotive applications.
AI offers opportunities to advance fusion energy R&D but requires responsible practices and expert collaborations to overcome its inherent challenges.
citing papers explorer
-
SLE-FNO: Single-Layer Extensions for Task-Agnostic Continual Learning in Fourier Neural Operators
SLE-FNO achieves zero forgetting and strong plasticity-stability balance in continual learning for FNO surrogate models of pulsatile blood flow by adding minimal single-layer extensions across four out-of-distribution tasks.
-
Robust and Explainable 3D Mode Shape Recognition Using Region-Aware Graph Neural Networks
A canonical engineering graph representation combined with region-aware graph attention learning enables robust and transferable 3D mode shape classification across heterogeneous vehicle models and sensor layouts.
-
A Hybridizable Neural Time Integrator for Stable Autoregressive Forecasting
A hybrid transformer-FEM integrator provides provable discrete energy preservation and gradient bounds for stable autoregressive forecasting of chaotic systems, with 65x fewer parameters and 9000x speedup in a fusion surrogate trained on 12 simulations.
-
Landsat-Sentinel-2 Algal Bloom Mapping Using Vision Transformers: Model Description, Implementation, and Examples
Vision transformers trained on a new global dataset of Landsat-Sentinel-2 patches detect floating coastal algal blooms with 8-65% omission/commission error and outperform spectral indices under cloud and glint conditions.
-
From inverse problems to neural operators: prediction, mechanism, and generalization of data-driven models
A conceptual argument that data-driven scientific models generalize outside their training data only when their mathematical form matches the true governing equation, which rules out neural operators and neural boundary-value models.
-
Component-Based Reduced-Order Modeling Framework for Rocket Combustion Dynamics in Multi-Injector Configurations
A component-based reduced-order modeling framework decomposes multi-injector rocket combustors into trainable sub-models that couple to predict combustion dynamics across flow and geometry changes.
-
ATHENA: Agentic Team for Hierarchical Evolutionary Numerical Algorithms
A multi-agent LLM framework, ATHENA, autonomously designs and refines numerical PDE solvers and physics-informed models, reportedly beating expert-authored baselines.
-
Advances in Scientific Machine Learning for Coupled Fluid Flow and Transport
Reviews linear and nonlinear SciML surrogates for coupled fluid flow and transport, with new PINN modeling of turbidity currents and β-VAE mode extraction from Rayleigh-Bénard convection.
-
Replay-Based Continual Learning for Physics-Informed Neural Operators
A replay-based continual learning strategy for physics-informed neural operators mitigates catastrophic forgetting on prior physical problems while enabling efficient adaptation to new data using only physical constraints.
-
Toward Generalizable Graph Learning for 3D Engineering AI: Explainable Workflows for CAE Mode Shape Classification and CFD Field Prediction
A graph learning framework turns heterogeneous 3D engineering data into physics-aware graphs processed by GNNs for CAE mode classification and CFD field prediction in automotive applications.
-
Challenges and opportunities for AI to help deliver fusion energy
AI offers opportunities to advance fusion energy R&D but requires responsible practices and expert collaborations to overcome its inherent challenges.