Self-attention mechanisms are used to build mesh-preserving neural surrogates that approximate PFEM dynamics for free-surface flows, delivering accurate transient predictions and improved scalability on 2D and 3D benchmarks.
arXiv preprint arXiv:2512.20399 , year=
8 Pith papers cite this work. Polarity classification is still indexing.
8
Pith papers citing it
abstract
We present GeoTransolver, a multiscale geometry-aware physics attention transformer for Computer Aided Engineering (CAE). GeoTransolver extends the Transolver backbone with GALE (Geometry-Aware Latent Embeddings) attention, which pairs physics-aware self-attention on learned state slices with cross-attention to a shared geometry and global context computed via multi-scale ball queries (inspired by Domino) and reused in every block. Implemented and released in NVIDIA PhysicsNeMo, GeoTransolver persistently projects geometry and global parameters, into physical state spaces to anchor computations to domain structure and operating regimes. We benchmark on DrivAerML, SHIFT-SUV, and SHIFT-Wing against Domino, Transolver (PhysicsNeMo implementation), and literature-reported AB-UPT, evaluating drag/lift R2 and relative L1 errors on field variables. As an additional nonlinear structural mechanics application, we also report Transolver and GeoTransolver results on bumper-beam and full-vehicle Body-in-White (BIW) crash-dynamics benchmarks, evaluating relative L2 trajectory error and probe-level kinematic MSE. GeoTransolver delivers improved accuracy, robustness to geometry and regime shifts, and favorable data efficiency; we include DrivAerML ablations and qualitative contour and design-trend results, advancing operator learning for high-fidelity surrogates on complex, irregular, non-linear domains.
citation-role summary
background 1
citation-polarity summary
years
2026 8roles
background 1polarities
background 1representative citing papers
ShardTensor is a domain-parallelism system for SciML that enables flexible scaling of extreme-resolution spatial datasets by removing the constraint of batch size one per device.
CarCrashNet supplies a large multi-modal crash simulation benchmark and CrashSolver neural model for data-driven full-vehicle crash prediction, validated against experiments and commercial solvers.
AeroJEPA applies joint-embedding predictive learning to produce scalable, semantically organized latent representations for 3D aerodynamic fields that support both field reconstruction and downstream design tasks.
LoRA adapters enable a 61.47M-parameter aerodynamics Transformer pretrained on four vehicle families to adapt to a held-out fifth family with 20 samples, reaching R²=0.85 and outperforming full fine-tuning and from-scratch training with 3x more data.
GeoTransolver applies geometry-aware operator learning and low-rank attention to predict high-fidelity crash dynamics on bumper and full-vehicle datasets, with one-shot temporal prediction achieving state-of-the-art accuracy and reduced overhead.
Hybrid mesh GNNs with geometry-aware attention achieve 3.20 mm temporal RMSE on a 25-sample full-vehicle lateral pole-impact test set while preserving interpretable displacement fields.
RETO achieves relative L2 errors of 0.063 on ShapeNet and 0.089/0.097 on DrivAerML surface pressure/velocity, outperforming Transolver and other baselines.
citing papers explorer
-
Attention mechanism for scalable mesh-based neural surrogates of free-surface fluids
Self-attention mechanisms are used to build mesh-preserving neural surrogates that approximate PFEM dynamics for free-surface flows, delivering accurate transient predictions and improved scalability on 2D and 3D benchmarks.
-
ShardTensor: Domain Parallelism for Scientific Machine Learning
ShardTensor is a domain-parallelism system for SciML that enables flexible scaling of extreme-resolution spatial datasets by removing the constraint of batch size one per device.
-
CarCrashNet: A Large-Scale Dataset and Hierarchical Neural Solver for Data-Driven Structural Crash Simulation
CarCrashNet supplies a large multi-modal crash simulation benchmark and CrashSolver neural model for data-driven full-vehicle crash prediction, validated against experiments and commercial solvers.
-
AeroJEPA: Learning Semantic Latent Representations for Scalable 3D Aerodynamic Field Modeling
AeroJEPA applies joint-embedding predictive learning to produce scalable, semantically organized latent representations for 3D aerodynamic fields that support both field reconstruction and downstream design tasks.
-
Adapting Automotive Aerodynamics Surrogates to New Vehicle Families via Transfer Learning
LoRA adapters enable a 61.47M-parameter aerodynamics Transformer pretrained on four vehicle families to adapt to a held-out fifth family with 20 samples, reaching R²=0.85 and outperforming full fine-tuning and from-scratch training with 3x more data.
-
High-Fidelity Industrial Crash Dynamics Prediction via Geometry-Aware Operator Learning with Memory-Efficient Low-Rank Attention
GeoTransolver applies geometry-aware operator learning and low-rank attention to predict high-fidelity crash dynamics on bumper and full-vehicle datasets, with one-shot temporal prediction achieving state-of-the-art accuracy and reduced overhead.
-
Crash Assessment via Mesh-Based Graph Neural Networks and Physics-Aware Attention
Hybrid mesh GNNs with geometry-aware attention achieve 3.20 mm temporal RMSE on a 25-sample full-vehicle lateral pole-impact test set while preserving interpretable displacement fields.
-
RETO: A Rotary-Enhanced Transformer Operator for High-Fidelity Prediction of Automotive Aerodynamics
RETO achieves relative L2 errors of 0.063 on ShapeNet and 0.089/0.097 on DrivAerML surface pressure/velocity, outperforming Transolver and other baselines.