Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Random Process Flow Matching: Generative Implicit Representations of Multivariate Random Fields
Julien Lalanne, David Picard, Lionel Boillot, Lina-María GUAYACÁN-CARRILLO, Leon Barens, Jean-Michel Pereira
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 RP Flow instead of calibrated GPR | image regression 4x upsampling · Div2K 4x upsampling | PCE1: 0.12 -> 0.09 | Improved |
| use RP Flow instead of the Rff Network | image regression · Div2K 4x upsampling | PSNR: 25.20 -> 25.30 | No clear effect |
| use RP Flow instead of noiseless GPR | image regression · Div2K 25% random pixels | PSNR: 17.94 -> 23.70 | Improved |
| use RP Flow M instead of Rff Network | seismic interpolation · Seismic interpolation (synthetic) | PSNR: 26.78 -> 32.08 | Improved |
| use RP Flow M instead of Rff Network | seismic interpolation · Seismic interpolation (real F3) | PSNR: 26.11 -> 28.77 | Improved |
| increase model size from S (1.7M) to M (10M) (ablation) | RP Flow · Seismic interpolation (synthetic) | PSNR: 31.24 -> 32.08 | Improved |
| increase model size from M (10M) to L (50M) (ablation) | RP Flow · Seismic interpolation (synthetic) | PSNR: 32.08 -> 31.84 | Hurt |
| increase model size from S (1.7M) to L (50M) (ablation) | RP Flow · Seismic interpolation (real F3) | PCE: 0.01 -> 0.04 | Hurt |
| reduce Euler steps k from 100 to 10 (ablation) | RP Flow · Seismic interpolation | computation time: factor 10 | No clear effect |
| compare RP Flow predictive variance against GPR at optimal calibration | RP Flow · Div2K upsampling | predictive variance: 0.16 -> 0.08 | Improved |
| use a discontinuous source process instead of continuous (ablation) | RP Flow · 1D discontinuous target | regularity | Improved |
Closest papers by meaning
- Flow Matching Neural Processes (NeurIPS 2025 · similarity 0.91)
- Flow Matching with Gaussian Process Priors for Probabilistic Time Series Forecasting (ICLR 2025 · similarity 0.88)
- On the Closed-Form of Flow Matching: Generalization Does Not Arise from Target Stochasticity (NeurIPS 2025 · similarity 0.87)
- Analysis of Learning a Flow-based Generative Model from Limited Sample Complexity (ICLR 2024 · similarity 0.87)
- Stream-level Flow Matching with Gaussian Processes (ICML 2025 · similarity 0.87)
- Discrete Flow Matching (NeurIPS 2024 · similarity 0.87)
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.
