Research map / Probabilistic and scientific modeling / Molecules and proteins
UniDrag: Unified Multi-Field Prediction and Robust Shape Optimization for Vehicle Aerodynamics
Ye Liu, Shouyi Liu, Ding Wang, Huiyu Yang, Ruizhe Deng, Qian Li, Yuxiao Hu, Jianghang gu, Yongzheng Liu, Quanshi Zhang, Shiyi Chen, Yuntian Chen
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add modality-protected learning (gradient isolation) (ablation) | UniDrag · UniDrag 15k vehicle test set | Cd R2: 0.898 -> 0.937 | Improved |
| add local geometry injection to PSA (ablation) | UniDrag · UniDrag 15k vehicle test set | Cd R2: 0.905 -> 0.937 | Improved |
| add 3D-RoPE on slice centers (ablation) | UniDrag · UniDrag 15k vehicle test set | Cd R2: 0.915 -> 0.937 | Improved |
| add gated expert routing (MoE-FFN) in last 4 blocks (ablation) | UniDrag · UniDrag 15k vehicle test set | Cd R2: 0.921 -> 0.937 | Improved |
| add QK-RMSNorm (ablation) | UniDrag · UniDrag 15k vehicle test set | Cd R2: 0.924 -> 0.937 | Improved |
| add modality-protected learning (ablation) | UniDrag · UniDrag 15k vehicle test set | Build-up R2: 0.945 -> 0.971 | Improved |
| raise EoT samples M from 1 to 5 (ablation) | Robust optimization · CFD verification, 50 samples | CFD Cd reduction / success rate: -11.2% -> -13.7%; 97% -> 100% | Improved |
| raise EoT samples M from 5 to 7 (ablation) | Robust optimization · CFD verification, 50 samples | CFD Cd reduction: -13.7% -> -13.5% | No clear effect |
| refine FFD grid from 5x5x5 to 9x9x9 (ablation) | FFD optimization · CFD verification, 50 samples | CFD Cd reduction: -13.7% -> -16.1% | Improved |
| coarsen FFD grid to 3x3x3 (ablation) | FFD optimization · CFD verification, 50 samples | CFD Cd reduction: -13.7% -> -6.8% | Hurt |
| replace GeoTransolver with UniDrag for Cd prediction | Surrogate · UniDrag 15k vehicle test set | Cd R2: 0.8710 -> 0.9370 | Improved |
| constrained FFD optimization instead of generative methods (TripOptimizer, PhysGen) | Shape optimization · Qualitative | geometry validity | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Transolver — Transolver: A Fast Transformer Solver for PDEs on General Geometries (ICML 2024) · better on 4 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- DrivAerStar: An Industrial-Grade CFD Dataset for Vehicle Aerodynamic Optimization (NeurIPS 2025 · similarity 0.88)
- Aerodynamic Coefficients Prediction via Cross-Attention Fusion and Physical-Informed Training (AAAI 2025 · similarity 0.88)
- 3DID: Direct 3D Inverse Design for Aerodynamics with Physics-Aware Optimization (NeurIPS 2025 · similarity 0.88)
- AeroGTO: An Efficient Graph-Transformer Operator for Learning Large-Scale Aerodynamics of 3D Vehicle Geometries (AAAI 2025 · similarity 0.88)
- DragSolver: A Multi-Scale Transformer for Real-World Automotive Drag Coefficient Estimation (ICML 2025 · similarity 0.87)
- DrivAerNet++: A Large-Scale Multimodal Car Dataset with Computational Fluid Dynamics Simulations and Deep Learning Benchmarks (NeurIPS 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
