Skip to content

Whack-a-mole Online Learning: Physics-Informed Neural Network for Intraday Implied Volatility Surface

Bibliographic record. Follow the original-source link for the publication.

Field Value
Primary domain Volatility
Other domains
Methods Financial Ml
Facets
Authors Kentaro Hoshisashi, Carolyn E. Phelan, Paolo Barucca
Published 2024-11-04
Source arXiv Quantitative Finance History
Identifiers arxiv:2411.02375
URL Open original source

Editorial synthesis

Why it matters

Under sparse intraday IV data, a method that enforces PDE and no-arbitrage constraints in real-time could reduce calibration latency and consistency risk, which is highly relevant for operational risk control. (abstract:S1, abstract:S2, abstract:S3, abstract:S4, abstract:S5, abstract:S6)

Main author claims

  • The authors claim WamOL introduces self-adaptive balancing of losses to enforce PDE and no-arbitrage constraints while fitting IV surfaces. (abstract:S3, abstract:S4)
  • They claim superior intraday IVS calibration from sparse data and improved capture of dynamic option-price/risk profile evolution. (abstract:S1, abstract:S5, abstract:S6)

Data, method, or discussion scope

The evidence scope is limited to abstract-level method framing and claimed experiments; no implementation details, windowing scheme, numerical stability checks, or benchmark numbers are given. (abstract:S1, abstract:S2, abstract:S3, abstract:S4, abstract:S5, abstract:S6)

Main limitations

The claim of superior performance lacks explicit metrics and comparator names. In production, opaque adaptive weighting could over-smooth or become unstable around market extremes. (abstract:S4, abstract:S5, abstract:S2)

Relationships

  • None recorded.