Research map / Probabilistic and scientific modeling / Dynamical systems
Learning Population-Level Diffusions with Generative RNNs
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 |
|---|---|---|---|
| stationary contrastive-divergence pre-training before backprop through time (ablation) | diffusive RNN potential · Himmelblau flow | training log-likelihood and runtime | Helped |
| sum-of-ramps RNN parametrization of the potential instead of local Gaussian model | population-level diffusion inference · Styblinski flow in high dimensions (d=50) | held-out Wasserstein distance | Helped |
| RNN parametrization vs. linear and Ornstein-Uhlenbeck (quadratic) parametric models | quadratic potential flow · Quadratic flow d=2,10,50 | held-out Wasserstein distance | No clear effect |
| RNN parametrization vs. local and parametric baselines | single-cell RNA-seq D4 prediction · top 5 and 10 differential genes | held-out Wasserstein distance | Helped |
| increasing number of genes from 10 to higher dimensions | RNN model on scRNA-seq · gene expression D4 prediction | prediction accuracy | Hurt |
Closest papers by meaning
- Learning diffusion at lightspeed (NeurIPS 2024 · similarity 0.87)
- Diffusion Models are Evolutionary Algorithms (ICLR 2025 · similarity 0.86)
- Tracing the Roots: Leveraging Temporal Dynamics in Diffusion Trajectories for Origin Attribution (NeurIPS 2025 · similarity 0.86)
- Reinforcement Learning for Control of Non-Markovian Cellular Population Dynamics (ICLR 2025 · similarity 0.85)
- Dynamical Diffusion: Learning Temporal Dynamics with Diffusion Models (ICLR 2025 · similarity 0.85)
- Learning stochastic dynamics from snapshots through regularized unbalanced optimal transport (ICLR 2025 · 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.