Stochastic gradient descent in continuous time

DOI10.1137/17M1126825MaRDI QIDQ4607057zbMATH OpenOpenAlexFDO

Authors Justin A. Sirignano, Konstantinos Spiliopoulos

Publication date 12 March 2018

Published in SIAM Journal on Financial Mathematics (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1611.05545

zbMATH Keywords

American optionsdeep learningmachine learningstatistical learningstochastic differential equationsstochastic gradient descent

Mathematics Subject Classification ID

Martingales with continuous parameter (60G44) Stochastic ordinary differential equations (aspects of stochastic analysis) (60H10) Neural nets and related approaches to inference from stochastic processes (62M45) Learning and adaptive systems in artificial intelligence (68T05) Derivative securities (option pricing, hedging, etc.) (91G20)

Abstract: Stochastic gradient descent in continuous time (SGDCT) provides a computationally efficient method for the statistical learning of continuous-time models, which are widely used in science, engineering, and finance. The SGDCT algorithm follows a (noisy) descent direction along a continuous stream of data. SGDCT performs an online parameter update in continuous time, with the parameter updates $het{a}_t$ satisfying a stochastic differential equation. We prove that $$ where $$ is a natural objective function for the estimation of the continuous-time dynamics. The convergence proof leverages ergodicity by using an appropriate Poisson equation to help describe the evolution of the parameters for large times. SGDCT can also be used to solve continuous-time optimization problems, such as American options. For certain continuous-time problems, SGDCT has some promising advantages compared to a traditional stochastic gradient descent algorithm. As an example application, SGDCT is combined with a deep neural network to price high-dimensional American options (up to 100 dimensions).

Recommendations

• Stochastic gradient descent in continuous time: a central limit theorem
• Analysis of stochastic gradient descent in continuous time
• Application of kernel-based stochastic gradient algorithms to option pricing
• An improvement of stochastic gradient descent approach for mean-variance portfolio optimization problem
• scientific article; zbMATH DE number 2133131

Cites work

• An equilibrium characterization of the term structure
• Asymptotic inference for stochastic processes
• Closed-form likelihood expansions for multivariate diffusions
• Convergence of numerical schemes for the solution of parabolic stochastic partial differential equations
• Deep learning
• DGM: a deep learning algorithm for solving partial differential equations
• Exponential convergence of Langevin distributions and their discrete approximations
• Gradient Convergence in Gradient methods with Errors
• High order discretization schemes for the CIR process: application to affine term structure and heston models
• scientific article; zbMATH DE number 48727 (Why is no real title available?)
• scientific article; zbMATH DE number 1972910 (Why is no real title available?)
• scientific article; zbMATH DE number 1432782 (Why is no real title available?)
• scientific article; zbMATH DE number 5066253 (Why is no real title available?)
• Likelihood Inference for Discretely Observed Nonlinear Diffusions
• Maximum Likelihood Estimation of Discretely Sampled Diffusions: A Closed-form Approximation Approach
• On Poisson equation and diffusion approximation. II.
• On the Poisson equation and diffusion approximation. I
• OPTIMAL MEAN REVERSION TRADING WITH TRANSACTION COSTS AND STOP-LOSS EXIT
• Parameter estimation in stochastic differential equations.
• Pricing of foreign exchange options under the Heston stochastic volatility model and CIR interest rates
• Robust Stochastic Approximation Approach to Stochastic Programming
• SOLUTION OF THE EXTENDED CIR TERM STRUCTURE AND BOND OPTION VALUATION
• Speculative futures trading under mean reversion
• Statistical inference for ergodic diffusion processes.

Cited in

(32)

• Stochastic gradient algorithm with random truncations
• Online drift estimation for jump-diffusion processes
• An improvement of stochastic gradient descent approach for mean-variance portfolio optimization problem
• Analysis of stochastic gradient descent in continuous time
• Perturbed iterate SGD for Lipschitz continuous loss functions
• Online adjoint methods for optimization of PDEs
• Continuous-time gradient-like descent algorithm for constrained convex unknown functions: penalty method application
• The deep learning Galerkin method for the general Stokes equations
• Two-timescale stochastic gradient descent in continuous time with applications to joint online parameter estimation and optimal sensor placement
• Discontinuous gradient algorithm for finite-time estimation of time-varying parameters
• Deep-Learning Solution to Portfolio Selection with Serially Dependent Returns
• On the fast convergence of random perturbations of the gradient flow
• Solving high-dimensional optimal stopping problems using deep learning
• Trading signals in VIX futures
• Mean Field Analysis of Deep Neural Networks
• Stochastic gradient descent in continuous time: a central limit theorem
• Gradient-Based Discrete-Time Concurrent Learning for Standalone Function Approximation
• Asymptotic analysis via stochastic differential equations of gradient descent algorithms in statistical and computational paradigms
• Adaptive robust control in continuous time
• Deep learning for limit order books
• Deep learning: an introduction for applied mathematicians
• Joint online parameter estimation and optimal sensor placement for the partially observed stochastic advection-diffusion equation
• Online parameter estimation for the McKean-Vlasov stochastic differential equation
• Stochastic gradient descent with noise of machine learning type. II: Continuous time analysis
• Continuous‐time stochastic gradient descent for optimizing over the stationary distribution of stochastic differential equations
• The Continuous Stochastic Gradient Method: Part I -- Convergence Theory
• Value iteration for streaming data on a continuous space with gradient method in an RKHS
• Stochastic gradient descent in continuous time for drift identification in multiscale diffusions
• Parametric estimation of stochastic differential equations via online gradient descent
• Filtered data based estimators for stochastic processes driven by colored noise
• OGF: an online gradient flow method for optimizing the statistical steady-state time averages of unsteady turbulent flows
• Learning two parameters in fractional Tikhonov regularization method via deep neural networks: application to the inverse problem for the Laplace operator with noisy data

Describes a project that uses

Uses Software

• DGM

This page was built for publication: Stochastic gradient descent in continuous time

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4607057)