Machine Learning Approximation Algorithms for High-Dimensional Fully Nonlinear Partial Differential Equations and Second-order Backward Stochastic Differential Equations

High-dimensional partial differential equations (PDEs) appear in a number of models from the financial industry, such as in derivative pricing models, credit valuation adjustment models, or portfolio optimization models. The PDEs in such applications are high-dimensional as the dimension corresponds to the number of financial assets in a portfolio. Moreover, such PDEs are often fully nonlinear due to the need to incorporate certain nonlinear phenomena in the model such as default risks, transaction costs, volatility uncertainty (Knightian uncertainty), or trading constraints in the model. Such high-dimensional fully nonlinear PDEs are exceedingly difficult to solve as the computational effort for standard approximation methods grows exponentially with the dimension. In this work, we propose a new method for solving high-dimensional fully nonlinear second-order PDEs. Our method can in particular be used to sample from high-dimensional nonlinear expectations. The method is based on (1) a connection between fully nonlinear second-order PDEs and second-order backward stochastic differential equations (2BSDEs), (2) a merged formulation of the PDE and the 2BSDE problem, (3) a temporal forward discretization of the 2BSDE and a spatial approximation via deep neural nets, and (4) a stochastic gradient descent-type optimization procedure. Numerical results obtained using TensorFlow in Python illustrate the efficiency and the accuracy of the method in the cases of a 100-dimensional Black–Scholes–Barenblatt equation, a 100-dimensional Hamilton–Jacobi–Bellman equation, and a nonlinear expectation of a 100-dimensional G-Brownian motion.

[1]  Cornelis W. Oosterlee,et al.  A Fourier-Cosine Method for an Efficient Computation of Solutions to BSDEs , 2013, SIAM J. Sci. Comput..

[2]  G. Pagès,et al.  A quantization algorithm for solving multidimensional discrete-time optimal stopping problems , 2003 .

[3]  Hai-ping Shi.  Backward stochastic differential equations in finance , 2010 .

[4]  Tara N. Sainath,et al.  Deep Neural Networks for Acoustic Modeling in Speech Recognition: The Shared Views of Four Research Groups , 2012, IEEE Signal Processing Magazine.

[5]  Christian Bender,et al.  A PRIMAL–DUAL ALGORITHM FOR BSDES , 2013, 1310.3694.

[6]  Justin A. Sirignano,et al.  DGM: A deep learning algorithm for solving partial differential equations , 2017, J. Comput. Phys..

[7]  H. McKean.  Application of brownian motion to the equation of kolmogorov-petrovskii-piskunov , 1975 .

[8]  E Weinan,et al.  Overcoming the curse of dimensionality: Solving high-dimensional partial differential equations using deep learning , 2017, ArXiv.

[9]  E. Tadmor.  A review of numerical methods for nonlinear partial differential equations , 2012 .

[10]  M. Avellaneda,et al.  Pricing and hedging derivative securities in markets with uncertain volatilities , 1995 .

[11]  Pierre Henry-Labordere,et al.  Uncertain Volatility Model: A Monte-Carlo Approach , 2010 .

[12]  Shige Peng,et al.  NONLINEAR EXPECTATIONS AND NONLINEAR MARKOV CHAINS , 2005 .

[13]  Sergey Ioffe,et al.  Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift , 2015, ICML.

[14]  G. N. Milstein,et al.  Numerical Algorithms for Forward-Backward Stochastic Differential Equations , 2006, SIAM J. Sci. Comput..

[15]  E Weinan,et al.  Deep Learning-Based Numerical Methods for High-Dimensional Parabolic Partial Differential Equations and Backward Stochastic Differential Equations , 2017, Communications in Mathematics and Statistics.

[16]  Klaus-Robert Müller,et al.  Efficient BackProp , 2012, Neural Networks: Tricks of the Trade.

[17]  Yoshua Bengio,et al.  Deep Sparse Rectifier Neural Networks , 2011, AISTATS.

[18]  C. Schwab,et al.  NUMERICAL SOLUTION OF PARABOLIC EQUATIONS IN HIGH DIMENSIONS , 2004 .

[19]  Céline Labart,et al.  A parallel algorithm for solving BSDEs , 2013, Monte Carlo Methods Appl..

[20]  Geoffrey E. Hinton,et al.  Deep Learning , 2015, Nature.

[21]  Plamen Turkedjiev,et al.  Two algorithms for the discrete time approximation of Markovian backward stochastic differential equations under local conditions , 2013, 1309.4378.

[22]  S. Osher,et al.  Algorithms for overcoming the curse of dimensionality for certain Hamilton–Jacobi equations arising in control theory and elsewhere , 2016, Research in the Mathematical Sciences.

[23]  Bruno Bouchard,et al.  Discrete-Time Approximation of BSDEs and Probabilistic Schemes for Fully Nonlinear PDEs , 2009 .

[24]  Emmanuel Gobet,et al.  Linear regression MDP scheme for discrete backward stochastic differential equations under general conditions , 2015, Math. Comput..

[25]  Andrew R. Barron,et al.  Universal approximation bounds for superpositions of a sigmoidal function , 1993, IEEE Trans. Inf. Theory.

[26]  Demis Hassabis,et al.  Mastering the game of Go with deep neural networks and tree search , 2016, Nature.

[27]  P. Forsyth,et al.  Implicit solution of uncertain volatility/transaction cost option pricing models with discretely observed barriers , 2001 .

[28]  Cornelis W. Oosterlee,et al.  Numerical Fourier Method and Second-Order Taylor Scheme for Backward SDEs in Finance , 2014 .

[29]  Stefan Geiss,et al.  Decoupling on the Wiener Space, Related Besov Spaces, and Applications to BSDEs , 2014, Memoirs of the American Mathematical Society.

[30]  Bruno Bouchard,et al.  Lecture notes on BSDEs Main existence and stability results , 2015 .

[31]  Jie Xiong,et al.  A branching particle system approximation for a class of FBSDEs , 2016 .

[32]  Shinzo Watanabe,et al.  On the branching process for Brownian particles with an absorbing boundary , 1965 .

[33]  Shige Peng,et al.  Probabilistic interpretation for systems of quasilinear parabolic partial differential equations , 1991 .

[34]  Jinguo Liu,et al.  Approximating quantum many-body wave functions using artificial neural networks , 2017, 1704.05148.

[35]  B. Øksendal.  Stochastic differential equations : an introduction with applications , 1987 .

[36]  Erhan Bayraktar,et al.  Valuation of Mortality Risk via the Instantaneous Sharpe Ratio: Applications to Life Annuities , 2008, 0802.3250.

[37]  P. Kloeden,et al.  Numerical Solution of Stochastic Differential Equations , 1992 .

[38]  J. Yong,et al.  Solving forward-backward stochastic differential equations explicitly — a four step scheme , 1994 .

[39]  S. Peng.  Nonlinear Expectations, Nonlinear Evaluations and Risk Measures , 2004 .

[40]  Johan A. K. Suykens,et al.  Learning solutions to partial differential equations using LS-SVM , 2015, Neurocomputing.

[41]  Michael V. Tretyakov,et al.  Discretization of forward–backward stochastic differential equations and related quasi-linear parabolic equations , 2007 .

[42]  Xiaolu Tan,et al.  A Numerical Algorithm for a Class of BSDE Via Branching Process , 2013 .

[43]  S. Shreve,et al.  Stochastic differential equations , 1955, Mathematical Proceedings of the Cambridge Philosophical Society.

[44]  Robert Denk,et al.  A forward scheme for backward SDEs , 2007 .

[45]  J. Bismut.  Conjugate convex functions in optimal stochastic control , 1973 .

[46]  G. Maruyama.  Continuous Markov processes and stochastic equations , 1955 .

[47]  Céline Labart,et al.  Solving BSDE with Adaptive Control Variate , 2010, SIAM J. Numer. Anal..

[48]  Nizar Touzi,et al.  A Probabilistic Numerical Method for Fully Nonlinear Parabolic PDEs , 2009, 0905.1863.

[49]  Arnaud Lionnet,et al.  Time discretization of FBSDE with polynomial growth drivers and reaction-diffusion PDEs , 2013, 1309.2865.

[50]  G. N. Mil’shtejn.  Approximate Integration of Stochastic Differential Equations , 1975 .

[51]  A. Ruszczynski,et al.  A Dual Method For Backward Stochastic Differential Equations with Application to Risk Valuation , 2017, 1701.06234.

[52]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[53]  Sebastian Ruder,et al.  An overview of gradient descent optimization algorithms , 2016, Vestnik komp'iuternykh i informatsionnykh tekhnologii.

[54]  Weidong Zhao,et al.  Efficient spectral sparse grid approximations for solving multi-dimensional forward backward SDEs , 2016 .

[55]  Arnulf Jentzen,et al.  A proof that artificial neural networks overcome the curse of dimensionality in the numerical approximation of Black-Scholes partial differential equations , 2018, Memoirs of the American Mathematical Society.

[56]  J. Muhle‐Karbe,et al.  Portfolio choice with small temporary and transient price impact , 2017, Mathematical Finance.

[57]  Jianfeng Zhang.  A numerical scheme for BSDEs , 2004 .

[58]  Michael Mascagni,et al.  Monte Carlo solution of Cauchy problem for a nonlinear parabolic equation , 2010, Math. Comput. Simul..

[59]  W. Zhao,et al.  High order numerical schemes for second-order FBSDEs with applications to stochastic optimal control , 2015, 1502.03206.

[60]  Geoffrey E. Hinton,et al.  Learning internal representations by error propagation , 1986 .

[61]  Jean-François Chassagneux,et al.  Numerical Stability Analysis of the Euler Scheme for BSDEs , 2014, SIAM J. Numer. Anal..

[62]  J. Douglas,et al.  Numerical methods for forward-backward stochastic differential equations , 1996 .

[63]  Emmanuel Gobet,et al.  Approximation of backward stochastic differential equations using Malliavin weights and least-squares regression , 2016, 1601.01186.

[64]  Matthias Troyer,et al.  Solving the quantum many-body problem with artificial neural networks , 2016, Science.

[65]  Dimitrios I. Fotiadis,et al.  Artificial neural networks for solving ordinary and partial differential equations , 1997, IEEE Trans. Neural Networks.

[66]  N. Touzi,et al.  On the Monte Carlo simulation of BSDEs: An improvement on the Malliavin weights , 2010 .

[67]  P. Briand,et al.  Simulation of BSDEs by Wiener Chaos Expansion , 2012, 1204.4137.

[68]  Emmanuel Gobet,et al.  Numerical simulation of BSDEs using empirical regression methods: theory and practice , 2005 .

[69]  Cornelis W. Oosterlee,et al.  Computational and Applied , 2015 .

[70]  Marc L. Ross.  Counterparty Risk and Funding: The Four Wings of the TVA , 2013 .

[71]  P. Protter,et al.  Numberical Method for Backward Stochastic Differential Equations , 2002 .

[72]  Jia Zhuo,et al.  A monotone scheme for high-dimensional fully nonlinear PDEs , 2012, 1212.0466.

[73]  M. Gunzburger,et al.  A SPARSE-GRID METHOD FOR MULTI-DIMENSIONAL BACKWARD STOCHASTIC DIFFERENTIAL EQUATIONS * , 2013 .

[74]  Yaacov Z. Bergman.  Option Pricing with Differential Interest Rates , 1995 .

[75]  Lexing Ying,et al.  Solving parametric PDE problems with artificial neural networks , 2017, European Journal of Applied Mathematics.

[76]  Dan Crisan,et al.  Solving Backward Stochastic Differential Equations Using the Cubature Method: Application to Nonlinear Pricing , 2010, SIAM J. Financial Math..

[77]  Kurt Hornik,et al.  Multilayer feedforward networks are universal approximators , 1989, Neural Networks.

[78]  Hyuk Lee,et al.  Neural algorithm for solving differential equations , 1990 .

[79]  Ioannis Karatzas,et al.  Brownian Motion and Stochastic Calculus , 1987 .

[80]  S. Menozzi,et al.  A Forward-Backward Stochastic Algorithm For Quasi-Linear PDEs , 2006, math/0603250.

[81]  Nadia Oudjane,et al.  Branching diffusion representation of semilinear PDEs and Monte Carlo approximation , 2016, Annales de l'Institut Henri Poincaré, Probabilités et Statistiques.

[82]  Dan Crisan,et al.  Probabilistic methods for semilinear partial differential equations. Applications to finance , 2010 .

[83]  Satish S. Udpa,et al.  Finite-element neural networks for solving differential equations , 2005, IEEE Transactions on Neural Networks.

[84]  W. E,et al.  Multilevel Picard iterations for solving smooth semilinear parabolic heat equations , 2016, Partial Differential Equations and Applications.

[85]  X. Warin.  Variations on branching methods for non linear PDEs , 2017, 1701.07660.

[86]  Pierre Henry-Labordere,et al.  Counterparty Risk Valuation: A Marked Branching Diffusion Approach , 2012, 1203.2369.

[87]  Philipp Petersen,et al.  Optimal approximation of piecewise smooth functions using deep ReLU neural networks , 2017, Neural Networks.

[88]  M. Nourian,et al.  Numerical solution of Helmholtz equation by the modified Hopfield finite difference techniques , 2009 .

[89]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[90]  H. Soner,et al.  Second‐order backward stochastic differential equations and fully nonlinear parabolic PDEs , 2005, math/0509295.

[91]  S. Peng,et al.  Backward stochastic differential equations and quasilinear parabolic partial differential equations , 1992 .

[92]  S. Peng,et al.  Adapted solution of a backward stochastic differential equation , 1990 .

[93]  Yoshua Bengio,et al.  Gradient-based learning applied to document recognition , 1998, Proc. IEEE.

[94]  E Weinan,et al.  On Multilevel Picard Numerical Approximations for High-Dimensional Nonlinear Parabolic Partial Differential Equations and High-Dimensional Nonlinear Backward Stochastic Differential Equations , 2017, Journal of Scientific Computing.

[95]  E. Gobet,et al.  A regression-based Monte Carlo method to solve backward stochastic differential equations , 2005, math/0508491.

[96]  Peter A. Forsyth,et al.  Hedging with a correlated asset: Solution of a nonlinear pricing PDE , 2007 .

[97]  Carlos Vázquez,et al.  Stratified Regression Monte-Carlo Scheme for Semilinear PDEs and BSDEs with Large Scale Parallelization on GPUs , 2016, SIAM J. Sci. Comput..

[98]  E. Bayraktar,et al.  Pricing options in incomplete equity markets via the instantaneous Sharpe ratio , 2007, math/0701650.

[99]  Anna Lisa Amadori,et al.  Nonlinear integro-differential evolution problems arising in option pricing: a viscosity solutions approach , 2003, Differential and Integral Equations.

[100]  Andrew J. Meade,et al.  The numerical solution of linear ordinary differential equations by feedforward neural networks , 1994 .

[101]  George Cybenko,et al.  Approximation by superpositions of a sigmoidal function , 1992, Math. Control. Signals Syst..

[102]  E Weinan,et al.  Deep Learning Approximation for Stochastic Control Problems , 2016, ArXiv.

[103]  Pratima Hebbar,et al.  Branching diffusion processes in periodic media , 2019 .

[104]  Philippe von Wurstemberger,et al.  Strong error analysis for stochastic gradient descent optimization algorithms , 2018, 1801.09324.

[105]  H. Leland..  Option Pricing and Replication with Transactions Costs , 1985 .

[106]  Johannes Muhle-Karbe,et al.  TRADING WITH SMALL PRICE IMPACT , 2014, 1402.5304.

[107]  P. Ma.  Forward Backward Stochastic Differential Equations , 2003 .

[108]  J. Chassagneux,et al.  Numerical simulation of quadratic BSDEs , 2013, 1307.5741.

[109]  H. Pham.  Feynman-Kac Representation of Fully Nonlinear PDEs and Applications , 2014, 1409.0625.

[110]  S. Peng.  Nonlinear Expectations and Stochastic Calculus under Uncertainty , 2010, Probability Theory and Stochastic Modelling.

[111]  Christel Geiss,et al.  Simulation of BSDEs with jumps by Wiener Chaos expansion , 2015, 1502.05649.

[112]  H. Soner,et al.  National Centre of Competence in Research Financial Valuation and Risk Management Working Paper No . 845 Homogenization and Asymptotics for Small Transaction Costs : The Multidimensional Case , 2013 .

[113]  E. Gobet,et al.  Rate of convergence of an empirical regression method for solving generalized backward stochastic differential equations , 2006 .

[114]  Dan Crisan,et al.  RUNGE-KUTTA SCHEMES FOR BACKWARD STOCHASTIC DIFFERENTIAL EQUATIONS , 2014 .

[115]  Jean-François Chassagneux,et al.  Linear Multistep Schemes for BSDEs , 2014, SIAM J. Numer. Anal..

[116]  D. Crisan,et al.  Second order discretization of backward SDEs and simulation with the cubature method , 2014 .

[117]  É. Pardoux,et al.  Forward-backward stochastic differential equations and quasilinear parabolic PDEs , 1999 .

[118]  Weidong Zhao,et al.  Probabilistic High Order Numerical Schemes for Fully Nonlinear Parabolic PDEs , 2015 .

[119]  B. Bouchard,et al.  Discrete-time approximation and Monte-Carlo simulation of backward stochastic differential equations , 2004 .

[120]  Jean-Paul Laurent,et al.  An overview of the valuation of collateralized derivative contracts , 2014 .

[121]  Arnulf Jentzen,et al.  Solving high-dimensional partial differential equations using deep learning , 2017, Proceedings of the National Academy of Sciences.

[122]  B. Øksendal.  Stochastic Differential Equations , 1985 .

[123]  S. Peng.  G -Expectation, G -Brownian Motion and Related Stochastic Calculus of Itô Type , 2006, math/0601035.

[124]  Yoshua. Bengio,et al.  Learning Deep Architectures for AI , 2007, Found. Trends Mach. Learn..