Diffusion models for dynamic volatility surface generation and data-driven hedging
Updated
Updated · arxiv.org · Sep 15
Diffusion models for dynamic volatility surface generation and data-driven hedging
1 articles · Updated · arxiv.org · Sep 15
Summary
Researchers have developed adaptive sequential diffusion models for generating dynamic implied-volatility surfaces in financial markets.
The models, AD-Seq-Vol and AD-Seq-Vol-FT, capture both cross-sectional and temporal dependencies and incorporate no-arbitrage constraints for realistic scenario generation.
Empirical tests on SPX options show these models reduce arbitrage violations and improve hedging performance, especially during periods of market stress like COVID-19.