Research map / Probabilistic and scientific modeling / Differential equations and PDEs
Self-Consistent Velocity Matching of Probability Flows
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 SCVM-NODE instead of adjoint method ADJ | NODE velocity training · Mixture of Gaussians d=10 | training time (minutes): 585.2 -> 7.37 | Helped |
| use SCVM-NODE instead of ADJ | NODE velocity training · Mixture of Gaussians | self-consistency decay | Helped |
| use SCVM-NODE instead of JKO-ICNN | gradient flow solver · Mixture of Gaussians d=60 | training time (minutes): 375.2 -> 23.9 | Helped |
| use SCVM-NODE instead of JKO methods | gradient flow solver · Mixture of Gaussians | symmetric KL | Helped |
| use ADJ instead of SCVM-TIPF/NODE | solver · Ornstein-Uhlenbeck d=5, 10 | divergence metrics | Hurt |
| use SCVM-NODE instead of ADJ | solver · Ornstein-Uhlenbeck d=10 | training time (minutes): 341 -> 9 | Helped |
| use SCVM-TIPF instead of SCVM-NODE | SCVM · Ornstein-Uhlenbeck low dimension | probability divergences | Helped |
| use SCVM-TIPF instead of JKO-ICNN | solver · Porous medium equation d=6 | training time (minutes): 102 -> 21 | Helped |
| use SCVM-TIPF instead of JKO-ICNN | solver · Porous medium equation d>3 | TV distance and symmetric f-divergence | Helped |
| vary DFE forward Euler step size | DFE · Ornstein-Uhlenbeck d=60 | Bures-Wasserstein distance: 28.11, 0.31, 0.46, 9.21 | Mixed |
| use SCVM instead of DFE | solver · Time-dependent Fokker-Planck | divergences | Helped |
| use fewer coupling layers with random masking in TIPF (ablation) | SCVM-TIPF · implementation | accuracy | Hurt |
Closest papers by meaning
- Mesh-Free Convolution: Learned Spectral Attenuation and Transport (NeurIPS 2026 · similarity 0.86)
- MC^2: Monte Carlo Correction for Fast Elliptic PDE Solving (NeurIPS 2026 · similarity 0.86)
- MultiPDENet: PDE-embedded Learning with Multi-time-stepping for Accelerated Flow Simulation (ICML 2025 · similarity 0.86)
- Simulation-Free Training of Neural ODEs on Paired Data (NeurIPS 2024 · similarity 0.86)
- CALM-PDE: Continuous and Adaptive Convolutions for Latent Space Modeling of Time-dependent PDEs (NeurIPS 2025 · similarity 0.85)
- First-Order Trajectory Matching: Fast Ensemble Predictions of Chaotic, Turbulent, Stochastic Systems (NeurIPS 2026 · similarity 0.85)
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.