Research map / Probabilistic and scientific modeling / Variational and generative sampling
Streaming Variational Monte Carlo
Yuan Zhao, Josue Nassar, Ian Jordan, Mónica F. Bugallo, Il Memming Park
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 SVMC online proposal adaptation instead of offline VSMC (100 particles) | VSMC · LDS | negative ELBO: 1195.9 -> 1188.3 | Improved |
| use SVMC instead of VSMC (1,000 particles) | VSMC · LDS | negative ELBO: 1183.6 -> 1178.3 | Improved |
| use SVMC instead of VSMC (10,000 particles) | VSMC · LDS | negative ELBO: 1179.8 -> 1173.8 | Improved |
| increase SVMC particles from 100 to 1,000 (ablation) | SVMC · LDS | negative ELBO: 1188.3 -> 1178.3 | Improved |
| increase SVMC particles from 1,000 to 10,000 (ablation) | SVMC · LDS | negative ELBO: 1178.3 -> 1173.8 | Improved |
| use SVMC with 100 particles instead of BPF with 125,000 | BPF · LDS | negative ELBO: 1177.0 -> 1188.3 | Hurt |
| use SVMC instead of VSMC in runtime | VSMC · LDS | runtime (s): 6390.2 -> 47.5 | Improved |
| use SVMC (200 particles) instead of BPF (10,000 particles) | BPF · Chaotic RNN | RMSE: .4 -> .34 | Improved |
| use SVMC instead of BPF | BPF · Chaotic RNN | negative ELBO: 24.16 -> 20.42 | Improved |
| use SVMC instead of UKF | UKF · Chaotic RNN | RMSE: 3.9 -> .34 | Improved |
| increase number of gradient steps per time step (ablation) | SVMC · Chaotic RNN | RMSE and ELBO | Improved |
| increase number of samples L used for gradient (ablation) | SVMC · Chaotic RNN | performance | Hurt |
Closest papers by meaning
- Online Variational Sequential Monte Carlo (ICML 2024 · similarity 0.90)
- StreamPhy: Streaming Inference of High-Dimensional Physical Dynamics via State Space Models (NeurIPS 2026 · similarity 0.89)
- eXponential FAmily Dynamical Systems (XFADS): Large-scale nonlinear Gaussian state-space modeling (NeurIPS 2024 · similarity 0.88)
- Predicting partially observable dynamical systems via diffusion models with a multiscale inference scheme (NeurIPS 2025 · similarity 0.88)
- Bayesian Online Natural Gradient (BONG) (NeurIPS 2024 · similarity 0.87)
- Variational Inference for SDEs Driven by Fractional Noise (ICLR 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.
