Research map / Time series / Causal and generative time series
Generative Modeling of Irregular Time Series via SDE-Induced Continuous-Discrete Variational Inference
Zexin Yuan, Qinliang Su, Junxi Xiao
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 |
|---|---|---|---|
| add EIP correlation parameterization (ablation) | SDE-VI without Complex/Flow · Physionet interpolation | test MSE (x10^-2): 0.094 -> 0.084 | Improved |
| add EIP correlation parameterization (ablation) | SDE-VI without Complex/Flow · USHCN extrapolation | test MSE (x10^-2): 0.709 -> 0.572 | Improved |
| add EIP to Flow variant (ablation) | SDE-VI Flow only · USHCN extrapolation | test MSE (x10^-2): 0.537 -> 0.485 | Improved |
| add EIP to Complex+Flow variant (ablation) | SDE-VI · Physionet interpolation | test MSE (x10^-2): 0.058 -> 0.050 | Improved |
| add conditional normalizing flows (ablation) | SDE-VI without EIP/Complex · Physionet interpolation | test MSE (x10^-2): 0.094 -> 0.075 | Improved |
| add complex-valued inference (ablation) | SDE-VI without EIP/Flow · Physionet interpolation | test MSE (x10^-2): 0.094 -> 0.084 | Improved |
| add complex-valued inference to EIP variant (ablation) | SDE-VI EIP only · USHCN extrapolation | test MSE (x10^-2): 0.572 -> 0.580 | No clear effect |
| add complex-valued inference to EIP+Flow (ablation) | SDE-VI · USHCN extrapolation | test MSE (x10^-2): 0.485 -> 0.455 | Improved |
| add Flow to EIP+Complex (ablation) | SDE-VI · Physionet interpolation | test MSE (x10^-2): 0.059 -> 0.050 | Improved |
| use SDE-VI instead of ACSSM | Irregular time series model · Physionet interpolation | test MSE (x10^-2): 0.116 -> 0.052 | Improved |
| use SDE-VI instead of ACSSM | Irregular time series model · USHCN interpolation | test MSE (x10^-2): 0.006 -> 0.044 | Hurt |
| use SDE-VI instead of ACSSM | Irregular time series model · USHCN extrapolation | test MSE (x10^-2): 0.941 -> 0.455 | Improved |
Closest papers by meaning
- SING: SDE Inference via Natural Gradients (NeurIPS 2025 · similarity 0.89)
- Closing the Approximation Gap in Simulation-free Latent SDEs (NeurIPS 2026 · similarity 0.88)
- Variational Inference for SDEs Driven by Fractional Noise (ICLR 2024 · similarity 0.88)
- In-Context Learning of Stochastic Differential Equations with Foundation Inference Models (NeurIPS 2025 · similarity 0.87)
- SDE Matching: Scalable and Simulation-Free Training of Latent Stochastic Differential Equations (ICML 2025 · similarity 0.87)
- Neural Stochastic Flows: Solver-Free Modelling and Inference for SDE Solutions (NeurIPS 2025 · 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.
