Finance-Informed Operator Learning for Option Pricing with Quantum-Compatible Realizations (Q7360686)

We propose finance-informed Deep Operator Networks (FI-DeepONet) for option pricing that combines Black-Scholes carrier and learned correction, enforcing no-arbitrage bounds using a smooth Financial Admissibility Layer. This approach achieves nearly an order of magnitude reduction in global relative price error versus vanilla and physics-informed baselines, with moderate OOD generalization on parameter-shift tests within the same local-volatility family.