Research map / Probabilistic and scientific modeling / Molecules and proteins
SUPA: A Lightweight Diagnostic Simulator for Machine Learning in Particle Physics
Atul Kumar Sinha, Daniele Paliotta, Bálint Máté, John Andrew Raine, Tobias Golling, François Fleuret
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 |
|---|---|---|---|
| use Transflowmer instead of PointFlow | point cloud generative model · SUPAv2 | mean W1 of point features eta, phi, r: 0.167 -> 0.029 | Improved |
| use Transflowmer instead of SetVAE | point cloud generative model · SUPAv3 | mean W1 of point features eta, phi, r: 0.510 -> 0.044 | Improved |
| use Transflowmer instead of PointFlow | point cloud generative model · SUPAv5 | mean W1 of point features eta, phi, r: 0.028 -> 0.042 | Hurt |
| use PointFlow instead of SetVAE | point cloud generative model · SUPAv2 | mean W1 of point feature E: 0.253 -> 0.462 | Hurt |
| use PointFlow instead of SetVAE | point cloud generative model · SUPAv3 | mean W1 of point features eta, phi, r: 0.510 -> 0.130 | Improved |
| use Transflowmer instead of PointFlow | point cloud generative model · SUPAv4 | mean W1 of point feature E: 0.459 -> 0.006 | Improved |
| use Transflowmer instead of PointFlow | point cloud generative model · SUPAv5 | mean W1 of point feature E: 0.090 -> 0.151 | Hurt |
| use Transflowmer instead of SetVAE | point cloud generative model · SUPAv1 | layer centroids: 16.244 -> 5.766 | Improved |
| use Transflowmer instead of PointFlow | point cloud generative model · SUPAv4 | layer energy: 12.895 -> 0.659 | Improved |
| train CaloFlow at 2x resolution instead of 1x | CaloFlow · SUPA high-resolution | mean discrepancy: 3.57 -> 6.78 | Hurt |
| train CaloFlow at 3x resolution instead of 1x | CaloFlow · SUPA high-resolution | mean discrepancy: 3.57 -> 8.29 | Hurt |
| test 1x data with a model trained at 3x resolution | CaloFlow · SUPA high-resolution | mean discrepancy: 3.57 -> 7.20 | Hurt |
Closest papers by meaning
- The Well: a Large-Scale Collection of Diverse Physics Simulations for Machine Learning (NeurIPS 2024 · similarity 0.83)
- Neural SPH: Improved Neural Modeling of Lagrangian Fluid Dynamics (ICML 2024 · similarity 0.83)
- Patch Hierarchical Attention Transformers for Efficient Particle Jet Tagging (NeurIPS 2026 · similarity 0.81)
- PDE-Transformer: Efficient and Versatile Transformers for Physics Simulations (ICML 2025 · similarity 0.81)
- Learning 3D Particle-based Simulators from RGB-D Videos (ICLR 2024 · similarity 0.81)
- Text2PDE: Latent Diffusion Models for Accessible Physics Simulation (ICLR 2025 · similarity 0.81)
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.
