Research map / Probabilistic and scientific modeling / Dynamical systems
Efficient and Flexible Inference for Stochastic Systems
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 |
|---|---|---|---|
| estimate states and parameters jointly with gradient matching instead of VGPA smoothing in an inner loop | Our approach · Lorenz63 D=3 | runtime: 31s -> 2.4s | Helped |
| same, on 40-dimensional Lorenz96 | Our approach · Lorenz96 D=40 | runtime: 6503s -> 14s | Helped |
| same, on 1000-dimensional Lorenz96 | Our approach · Lorenz96 D=1000 | runtime: 17345s -> 383s | Helped |
| increase diffusion variance by factor of 30 (ablation) | Our approach · Lorenz attractor parameter estimates | parameter accuracy: 1 -> 30 | Mixed |
| lengthen observed time interval from 10 to 60 secs (ablation) | Our approach · Lorenz attractor | parameter estimation accuracy: 10 -> 60 | Helped |
| lengthen observed time interval (ablation) | Our approach · Lorenz attractor, sigma parameter | estimation accuracy | No clear effect |
| compare with Archambeau et al. 2008 variational approach | Our approach · Lorenz96 D=40, true forcing 8 | parameter estimate accuracy | Helped |
| reduce observed variables to 1/3 (ablation) | Our approach · Lorenz96 D=40 | parameter estimate accuracy | Hurt |
| leave the y dimension unobserved (ablation) | Our approach · Lorenz attractor | trajectory recovery | No clear effect |
Closest papers by meaning
- Scalable Variational Inference for Dynamical Systems (NeurIPS 2017 · similarity 0.90)
- Closing the Approximation Gap in Simulation-free Latent SDEs (NeurIPS 2026 · similarity 0.89)
- ODIN: ODE-Informed Regression for Parameter and State Inference in Time-Continuous Dynamical Systems (AAAI 2020 · similarity 0.88)
- SING: SDE Inference via Natural Gradients (NeurIPS 2025 · similarity 0.86)
- Variational Inference for SDEs Driven by Fractional Noise (ICLR 2024 · similarity 0.85)
- Learning invariant representations of time-homogeneous stochastic dynamical systems (ICLR 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.