Universal approximation property of neural stochastic differential equations
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866910888125005824 |
|---|---|
| author | Kwossek, Anna P. Prömel, David J. Teichmann, Josef |
| author_facet | Kwossek, Anna P. Prömel, David J. Teichmann, Josef |
| contents | We identify various classes of neural networks that are able to approximate continuous functions locally uniformly subject to fixed global linear growth constraints. For such neural networks the associated neural stochastic differential equations can approximate general stochastic differential equations, both of Itô diffusion type, arbitrarily well. Moreover, quantitative error estimates are derived for stochastic differential equations with sufficiently regular coefficients. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_16696 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Universal approximation property of neural stochastic differential equations Kwossek, Anna P. Prömel, David J. Teichmann, Josef Probability Machine Learning Functional Analysis Mathematical Finance 41A29, 60H10, 68T07, 91G80 We identify various classes of neural networks that are able to approximate continuous functions locally uniformly subject to fixed global linear growth constraints. For such neural networks the associated neural stochastic differential equations can approximate general stochastic differential equations, both of Itô diffusion type, arbitrarily well. Moreover, quantitative error estimates are derived for stochastic differential equations with sufficiently regular coefficients. |
| title | Universal approximation property of neural stochastic differential equations |
| topic | Probability Machine Learning Functional Analysis Mathematical Finance 41A29, 60H10, 68T07, 91G80 |
| url | https://arxiv.org/abs/2503.16696 |