Research map / Image generation / GANs
Non-adversarial training of Neural SDEs with signature kernel scores
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 |
|---|---|---|---|
| use full signature kernel score instead of truncated (N=3) (ablation) | Neural SDE · gBm marginal t=57 | KS score: 0.2416 -> 0.1075 | Helped |
| use full signature kernel score instead of truncated (N=5) (ablation) | Neural SDE · rBergomi marginal t=6 | KS score: 0.1126 -> 0.1086 | Helped |
| use full signature kernel score instead of truncated (N=5) (ablation) | Neural SDE · EUR/USD marginal t=57 | KS score: 0.1608 -> 0.1049 | Helped |
| use full signature kernel score instead of truncated (N=5) (ablation) | Neural SDE · USD/JPY marginal t=57 | KS score: 0.2311 -> 0.1051 | Helped |
| use full signature kernel score instead of truncated (N=5) (ablation) | Neural SDE · EUR/USD marginal t=6 | KS score: 0.1098 -> 0.1270 | Mixed |
| replace adversarial SDE-GAN discriminator with signature kernel score | Neural SDE · rBergomi marginal t=6 | Type I error: 68.3% -> 5.4% | Helped |
| replace adversarial SDE-GAN discriminator with signature kernel score | Neural SDE · gBm marginal t=19 | Type I error: 5.2% -> 6.0% | Mixed |
| replace SDE-GAN with signature kernel score | Neural SDE · EUR/USD marginal t=57 | KS score: 0.4209 -> 0.1049 | Helped |
| replace SDE-GAN with signature kernel score | Neural SDE · gBm cross-correlation | MSE: 0.014688 -> 0.010718 | Helped |
| use full instead of truncated signature kernel score (ablation) | Neural SDE · gBm cross-correlation | MSE: 0.066745 -> 0.010718 | Helped |
| use full instead of truncated signature kernel score (ablation) | Neural SDE · rBergomi cross-correlation | MSE: 0.054731 -> 0.016785 | Helped |
| use full instead of truncated signature kernel score (ablation) | Neural SDE · USD/JPY cross-correlation | MSE: 0.036628 -> 0.003898 | Helped |
Closest papers by meaning
- Efficient Training of Neural Stochastic Differential Equations by Matching Finite Dimensional Distributions (ICLR 2025 · similarity 0.88)
- Neural SDEs as Infinite-Dimensional GANs (ICML 2021 · similarity 0.88)
- Generative Modeling of Regular and Irregular Time Series Data via Koopman VAEs (ICLR 2024 · similarity 0.84)
- SMaRt: Improving GANs with Score Matching Regularity (ICML 2024 · similarity 0.84)
- Neural Stochastic Flows: Solver-Free Modelling and Inference for SDE Solutions (NeurIPS 2025 · similarity 0.83)
- Generating Financial Time Series by Matching Random Convolutional Features (NeurIPS 2026 · similarity 0.83)
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.