AeroJEPA: Learning Semantic Latent Representations for Scalable 3D Aerodynamic Field Modeling

• URL: http://arxiv.org/abs/2605.05586v1
• Date: Thu, 07 May 2026 02:11:26 GMT
• Title: AeroJEPA: Learning Semantic Latent Representations for Scalable 3D Aerodynamic Field Modeling
• Abstract summary: We introduce AeroJEPA, a Predictive Architecture for aerodynamic field modeling.<n>Rather than predicting the full flow field directly from geometry, AeroJEPA predicts a target latent representation of the flow.<n>We show that the resulting latent space supports controlled probing, linear probing, concept-vector arithmetic, and a constrained design latent-optimization experiment.
• Score: 5.931951549967812
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Aerodynamic surrogate models are increasingly used to replace repeated high-fidelity CFD evaluations in many-query design settings, but current approaches still face two important limitations: they often scale poorly to the very large fields arising in realistic 3D aerodynamics, and they rarely produce latent representations that are directly useful for analysis and design. We introduce AeroJEPA, a Joint-Embedding Predictive Architecture for aerodynamic field modeling that addresses both issues. Rather than predicting the full flow field directly from geometry, AeroJEPA predicts a target latent representation of the flow from a context latent representation of the geometry and operating conditions, and optionally reconstructs the field through a continuous implicit decoder. This formulation decouples latent prediction from field resolution while encouraging the latent space to organize semantically. We evaluate AeroJEPA on two complementary datasets: HiLiftAeroML, which stresses the method in a high-fidelity regime with extremely large boundary-layer fields, and SuperWing, which tests large-scale generalization and latent-space optimization over a broad family of transonic wings. Across these benchmarks, AeroJEPA is competitive as a continuous surrogate for aerodynamic fields, scales naturally to high-resolution outputs, and learns context and predicted latents that encode geometry and aerodynamic quantities not used directly as supervision. We further show that the resulting latent space supports controlled interpolation, linear probing, concept-vector arithmetic, and a constrained design latent-optimization experiment. These results suggest that predictive latent learning is a promising direction for scalable and design-meaningful aerodynamic surrogate modeling.

Related papers

• Neural Field Ensembles for Aerodynamic Surface Prediction: Winning Solution to the ONERA CRM Wall Distribution 2025 Challenge [0.0]
  This paper presents the methodology that achieved first place in the ONERA CRM Wall Distribution Regression Challenge.<n>It focuses on predicting pressure and skin-friction coefficient distributions over the NASA Common Research Model wing-body-pylon-nacelle configuration.<n>The proposed approach formulates the problem as a conditional neural field mapping spatial coordinates, surface normals, and operating conditions to aerodynamic wall quantities.
  arXiv  Detail & Related papers  (2026-09-15T13:27:37Z)
• DESCENT: Directed Edge Scene Encoding for Airport Surface Movement Prediction [54.94692733670454]
  DESCENT is a transformer-based architecture designed to handle heterogeneous dynamics and strict topological constraints.<n>Our approach features a Potential Reachable Set (PRS) context sampling mechanism that adaptively collects airfield environment context.<n>Our evaluations on the Amelia-10 benchmark demonstrate significant performance improvements over state-of-the-art baselines.
  arXiv  Detail & Related papers  (2026-08-26T16:49:50Z)
• Full-field prediction for engineering-scale three-dimensional aircraft with multigrid-hierarchical learning [14.587446418342665]
  We propose MHLF, a multigrid-hierarchical learning framework for accelerating engineering-scale aircraft flow simulations.<n>Across three engineering-scale aircraft cases spanning Mach 0.15 to 6.0, MHLF accelerates convergence without sacrificing flow-field accuracy.
  arXiv  Detail & Related papers  (2026-05-26T09:13:16Z)
• Earth-o1: A Grid-free Observation-native Atmospheric World Model [81.55899748753434]
  We present Earth-o1, an observation-native atmospheric world model that overcomes structural limitations.<n>By integrating diverse sensor inputs into a unified, grid-free dynamical field, the model autonomously advances the atmospheric state in space and time.<n>We show that this fundamentally distinct paradigm enables direct, real-time forecasting and cross-sensor inference without the overhead of explicit numerical solvers.
  arXiv  Detail & Related papers  (2026-05-07T14:27:48Z)
• Faster by Design: Interactive Aerodynamics via Neural Surrogates Trained on Expert-Validated CFD [7.948688528462053]
  Computational Fluid Dynamics (CFD) is central to racecar aerodynamic development.<n>Its cost severely limits the design space exploration feasible within realistic budgets.<n>New AI-based surrogate models promise to alleviate this bottleneck.
  arXiv  Detail & Related papers  (2026-04-20T16:42:35Z)
• VGGT-World: Transforming VGGT into an Autoregressive Geometry World Model [59.789011777899965]
  VGGT-World is a geometry world model that side-steps video generation entirely and instead forecasts the temporal evolution of frozen geometry-foundation-model features.<n>We show that VGGT-World significantly outperforms the strongest baselines in depth forecasting while running 3.6-5 times faster with only 0.43B trainable parameters.
  arXiv  Detail & Related papers  (2026-03-13T04:56:43Z)
• Going with the Speed of Sound: Pushing Neural Surrogates into Highly-turbulent Transonic Regimes [19.286954413935025]
  We present a new dataset of CFD simulations for 3D wings in the transonic regime.<n>The dataset comprises volumetric and surface-level fields for around $30,000$ samples with unique geometry and inflow conditions.<n>We evaluate several state-of-the-art neural surrogates on our dataset, including Transolver and AB-UPT.
  arXiv  Detail & Related papers  (2025-11-26T15:06:19Z)
• Scaling Up Occupancy-centric Driving Scene Generation: Dataset and Method [54.461213497603154]
  Occupancy-centric methods have recently achieved state-of-the-art results by offering consistent conditioning across frames and modalities.<n>Nuplan-Occ is the largest occupancy dataset to date, constructed from the widely used Nuplan benchmark.<n>We develop a unified framework that jointly synthesizes high-quality occupancy, multi-view videos, and LiDAR point clouds.
  arXiv  Detail & Related papers  (2025-10-27T03:52:45Z)
• Towards scalable surrogate models based on Neural Fields for large scale aerodynamic simulations [0.0]
  This paper introduces a novel surrogate modeling framework for aerodynamic applications based on Neural Fields.<n>The proposed approach, MARIO, addresses non parametric geometric variability through an efficient shape encoding mechanism.<n>It enables training on significantly downsampled meshes, while maintaining consistent accuracy during full-resolution inference.
  arXiv  Detail & Related papers  (2025-05-14T10:49:08Z)
• C(NN)FD -- Deep Learning Modelling of Multi-Stage Axial Compressors Aerodynamics [0.0]
  This paper demonstrates the development and application of a generalized deep learning framework for predictions of the flow field and aerodynamic performance of multi-stage axial compressors.<n>A physics-based dimensionality reduction unlocks the potential for flow-field predictions for large-scale domains, re-formulating the regression problem from an unstructured to a structured one.<n>The proposed framework has the advantage of physically explainable predictions of overall performance, as the corresponding aerodynamic drivers can be identified on a 0D/1D/2D/3D level.
  arXiv  Detail & Related papers  (2025-03-18T15:58:58Z)
• An Efficient Occupancy World Model via Decoupled Dynamic Flow and Image-assisted Training [50.71892161377806]
  DFIT-OccWorld is an efficient 3D occupancy world model that leverages decoupled dynamic flow and image-assisted training strategy.<n>Our model forecasts future dynamic voxels by warping existing observations using voxel flow, whereas static voxels are easily obtained through pose transformation.
  arXiv  Detail & Related papers  (2024-12-18T12:10:33Z)
• A Geometry-Aware Message Passing Neural Network for Modeling Aerodynamics over Airfoils [61.60175086194333]
  aerodynamics is a key problem in aerospace engineering, often involving flows interacting with solid objects such as airfoils.<n>Here, we consider modeling of incompressible flows over solid objects, wherein geometric structures are a key factor in determining aerodynamics.<n>To effectively incorporate geometries, we propose a message passing scheme that efficiently and expressively integrates the airfoil shape with the mesh representation.<n>These design choices lead to a purely data-driven machine learning framework known as GeoMPNN, which won the Best Student Submission award at the NeurIPS 2024 ML4CFD Competition, placing 4th overall.
  arXiv  Detail & Related papers  (2024-12-12T16:05:39Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.