Paper: arXiv 2507.09412
Authors: Fabio Baschetti, Giacomo Bormetti, Pietro Rossi
Abstract
Joint calibration to SPX and VIX market data is a delicate task that requires sophisticated modeling and incurs significant computational costs. The latter is especially true when pricing of volatility derivatives hinges on nested Monte Carlo simulation. One such example is the 4-factor Markov Path-Dependent Volatility (PDV) model of Guyon and Lekeufack (2023). Nonetheless, its realism has earned it considerable attention in recent years. Gazzani and Guyon (2025) marked a relevant contribution by learning the VIX as a random variable, i.e., a measurable function of the model parameters and the Markovian factors. A neural network replaces the inner simulation and makes the joint calibration problem accessible. However, the minimization loop remains slow due to expensive outer simulation. The present paper overcomes this limitation by learning SPX implied volatilities, VIX futures, and VIX call option prices. The pricing functions reduce to simple matrix-vector products that can be evaluated on the fly, shrinking calibration times to just a few seconds.
Complexity vs Empirical Score
- Math Complexity: 8.5/10
- Empirical Rigor: 7.5/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: The paper employs advanced stochastic calculus, neural network theory, and regression techniques (high math), while also demonstrating practical implementation with specific calibration times, GPU usage, and a focus on making the model accessible for trading desks (high rigor).
Research Flowchart
flowchart TD
A["Research Goal<br>Joint SPX/VIX Calibration<br>to 4-Factor PDV Model"] --> B{"Data Inputs"}
B --> B1["SPX Market Data"]
B --> B2["VIX Market Data<br>Futures & Options"]
B --> C["Methodology<br>Deep Calibration via Neural Networks"]
C --> D["Computational Process<br>Forward Pass: Matrix-Vector Products"]
D --> E["Outcome 1<br>SPX Implied Volatilities"]
D --> F["Outcome 2<br>VIX Futures"]
D --> G["Outcome 3<br>VIX Call Option Prices"]
E & F & G --> H["Final Result<br>Joint Calibration in Seconds<br>vs. Nested Monte Carlo"]
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes neural networks that learn SPX implied volatilities, VIX futures and VIX call option prices, reducing pricing to matrix‑vector products and shrinking calibration times to a few seconds.
Source sentence
The present paper overcomes this limitation by learning SPX implied volatilities, VIX futures, and VIX call option prices. The pricing functions reduce to simple matrix-vector products that can be evaluated on the fly, shrinking calibration times to just a few seconds.
Assumes 5
- Volatility σt is assumed constant over each discretization interval of length δt. model
Source sentence
we assume volatility σt to be constant over each discretization interval of length δt
- The R2 components follow a simple ODE in continuous time. model
Source sentence
the R2 components follow a simple ODE in continuous time
- The quadruplet (R1,0,t, R1,1,t, R2,0,t, R2,1,t) is Markovian, enabling fast simulation. model
Source sentence
The choice of the two-exponential kernel makes the quadruplet (R1,0,t, R1,1,t, R2,0,t, R2,1,t) Markovian.
- Training data are generated by Monte Carlo simulation with a fixed number of trajectories. data
Source sentence
We fill the grid buckets using Monte Carlo simulation with Nsim = 218 trajectories.
Show 1 more
- Neural network inputs are model parameters θ and contract specifications (T, K); outputs are Black‑Scholes implied volatilities. model
Source sentence
A simple feed-forward neural network learns Black‑Scholes implied volatilities as a function of model parameters θ and contract specification (T, K).
Data
- Monte Carlo simulated SPX and VIX samples synthetic unknown
Source sentence
We fill the grid buckets using Monte Carlo simulation with Nsim = 218 trajectories.
Costs
Not applicable
Finds 2
- 95% of VIX futures pricing errors are below 7e‑04, confirming high accuracy of the approximation. predictive-accuracy
Source sentence
95% quantiles super-imposed in dashed red: q|err| 0.950 = 6.9774e-04
- Calibration of the joint SPX‑VIX problem can be performed in a matter of seconds. pricing-calibration
Source sentence
It now becomes a matter of seconds, which fact is essential for practitioners.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 11 related papers). Coverage is partial, so this is neither replication nor refutation.
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.