smq/directoryAn evolving quant resource archive
Back to directoryResource
Paper / 2025SMQ / 196902497491615
SigMA: Path Signatures and Multi-head Attention for Learning Parameters in fBm-driven SDEs
Chiheb Ben Hammouda / Cornelis W. Oosterlee / Xianglin Wu
deep learning / machine learningResource overview

SigMA: Path Signatures and Multi-head Attention for Learning Parameters in fBm-driven SDEs

Paper2025

by Chiheb Ben Hammouda, Cornelis W. Oosterlee, Xianglin Wu

Tags
deep learningmachine learningneural networksstatistical inferencestochastic differential equationsstochastic processes

Access resource

Open original source Understand this paper

Continue on SelfMadeQuant

Understand this paper

Suggested foundations for financial machine learning and stochastic processes and pricing. Specialist prerequisites vary by paper.

Before you read

  1. Time-Ordered Train, Validation and Test Design

    Preserve time order and avoid future information in training.

  2. Linear Regression

    Establish an interpretable baseline before a flexible model.

  3. Conditional Expectation

    Understand what it means to condition on current information.

  4. Stochastic Processes, Brownian Motion and Itô's Lemma

    Follow Brownian dynamics and the second-order term in Itô calculus.

Go further

  1. Finance-aware validation

    Account for dependence and overlapping observations.

  2. Feature importance

    Examine which information contributes to predictions.

  3. Market Model and Risk-Neutral Valuation

    Connect stochastic dynamics to no-arbitrage prices.

Explore a related reading path

Lesson links open public previews. Access to full lessons depends on the topic and your membership.

Explore Financial Machine Learning Explore Derivatives Pricing and Computational Finance

Keep exploring

Related resources

More ideas from similar topics.

6

From the same author

Chiheb Ben Hammouda, Cornelis W. Oosterlee, Xianglin Wu

6