Research map / Probabilistic and scientific modeling / Differential equations and PDEs
Distributional Gradient Matching for Learning Uncertain Neural Dynamics Models
Lenart Treven, Philippe Wenk, Florian Dörfler, Andreas Krause
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 DGM instead of SGHMC and SGLD on neural ODEs | neural ODE · Lotka Volterra 1 | log likelihood: 1.36 -> 1.96 | Improved |
| use DGM instead of SGHMC | neural ODE · Lorenz 1 | log likelihood: -3.02 -> -0.57 | Improved |
| use DGM instead of SGHMC | neural ODE · Quadrocopter 1 | log likelihood: -5.00 -> 0.64 | Improved |
| use DGM smoother for prediction instead of numerical integration of posterior samples | neural ODE · 100 test points | prediction time [ms]: 14.98 -> 0.68 | Improved |
| use DGM instead of NDP on multiple sparsely sampled trajectories | neural ODE · Lotka Volterra 100 | log likelihood: 0.62 -> 1.81 | Improved |
| use DGM instead of NDP | neural ODE · Double Pendulum 100 | log likelihood: 0.88 -> 1.86 | Improved |
| use DGM instead of NDP | neural ODE · Lorenz 125 | log likelihood: -2.85 -> -2.18 | Improved |
| increase number of training trajectories | NDP · Lotka Volterra | log likelihood | No clear effect |
| increase number of training trajectories | DGM · Lotka Volterra | log likelihood | Improved |
| increase model overparametrization (more matrix factors) | SGLD · linear system | log likelihood | Hurt |
| increase model overparametrization | SGHMC · linear system | log likelihood | Hurt |
| increase model overparametrization | DGM · linear system | log likelihood | No clear effect |
Closest papers by meaning
- Scalable Variational Inference for Dynamical Systems (NeurIPS 2017 · similarity 0.89)
- Simulation-Free Training of Neural ODEs on Paired Data (NeurIPS 2024 · similarity 0.86)
- How Deep Do We Need: Accelerating Training and Inference of Neural ODEs via Control Perspective (ICML 2024 · similarity 0.86)
- Efficient Training of Neural Stochastic Differential Equations by Matching Finite Dimensional Distributions (ICLR 2025 · similarity 0.86)
- Learning Physics Informed Neural ODEs with Partial Measurements (AAAI 2025 · similarity 0.86)
- From Fourier to Neural ODEs: Flow Matching for Modeling Complex Systems (ICML 2024 · 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.
