Paper: arXiv 2302.09176

Authors: Anastasis Kratsios, Cody Hyndman

Abstract

We consider the problem of simultaneously approximating the conditional distribution of market prices and their log returns with a single machine learning model. We show that an instance of the GDN model of Kratsios and Papon (2022) solves this problem without having prior assumptions on the market’s “clipped” log returns, other than that they follow a generalized Ornstein-Uhlenbeck process with a priori unknown dynamics. We provide universal approximation guarantees for these conditional distributions and contingent claims with a Lipschitz payoff function.

Complexity vs Empirical Score

  • Math Complexity: 8.5/10
  • Empirical Rigor: 2.5/10
  • Quadrant: Lab Rats — theoretically deep, empirically untested

Why this score: The paper presents rigorous mathematical proofs and universal approximation guarantees for generative models applied to generalized Ornstein-Uhlenbeck processes using geometric deep learning. However, it is purely theoretical with a complete lack of empirical backtesting or real-world market data validation.

Research Flowchart

  flowchart TD
    A[Goal: Jointly approximate market price & log-return conditional distributions] --> B[Inputs: Market prices & clipped log returns following unknown generalized OU process]
    B --> C[Method: Apply Geometric Deep Learning via GDN Model]
    C --> D[Compute conditional distribution & contingent claims with Lipschitz payoffs]
    D --> E[Outcome: Proven universal approximation guarantees without prior dynamic assumptions]