Convolutional Signature for Sequential Data

Signature is an infinite graded sequence of statistics known to characterize geometric rough paths, which includes the paths with bounded variation. This object has been studied successfully for machine learning with mostly applications in low dimensional cases. In the high dimensional case, it suffers from exponential growth in the number of features in truncated signature transform. We propose a novel neural network based model which borrows the idea from Convolutional Neural Network to address this problem. Our model reduces the number of features efficiently in a data dependent way. Some empirical experiments are provided to support our model.

[1]  Terry Lyons,et al.  The Signature of a Rough Path: Uniqueness , 2014, 1406.7871.

[2]  Patrick Kidger,et al.  Signatory: differentiable computations of the signature and logsignature transforms, on both CPU and GPU , 2020, ArXiv.

[3]  Franz J. Király,et al.  Kernels for sequentially ordered data , 2016, J. Mach. Learn. Res..

[4]  Ken-ichi Funahashi,et al.  On the approximate realization of continuous mappings by neural networks , 1989, Neural Networks.

[5]  R. Engle.  Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom inflation , 1982 .

[6]  Terry Lyons,et al.  Extracting information from the signature of a financial data stream , 2013, 1307.7244.

[7]  T. Bollerslev,et al.  Generalized autoregressive conditional heteroskedasticity , 1986 .

[8]  Imanol Perez Arribas,et al.  Non-parametric Pricing and Hedging of Exotic Derivatives , 2019, Applied Mathematical Finance.

[9]  Jeffrey Pennington,et al.  GloVe: Global Vectors for Word Representation , 2014, EMNLP.

[10]  J. Fouque,et al.  An Infinite-dimensional McKean-Vlasov Stochastic Equation , 2018 .

[11]  Yoshua Bengio,et al.  Learning Phrase Representations using RNN Encoder–Decoder for Statistical Machine Translation , 2014, EMNLP.

[12]  Imanol Perez Arribas.  Derivatives pricing using signature payoffs , 2018, 1809.09466.

[14]  Paris Perdikaris,et al.  Numerical Gaussian Processes for Time-Dependent and Nonlinear Partial Differential Equations , 2017, SIAM J. Sci. Comput..

[15]  Patrick Kidger,et al.  A Generalised Signature Method for Time Series , 2020, ArXiv.

[16]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[17]  Christopher Potts,et al.  Learning Word Vectors for Sentiment Analysis , 2011, ACL.

[18]  Terry Lyons,et al.  System Control and Rough Paths , 2003 .

[19]  Csaba Toth,et al.  Bayesian Learning from Sequential Data using Gaussian Processes with Signature Covariances , 2020, ICML.

[20]  Terry Lyons,et al.  A feature set for streams and an application to high-frequency financial tick data , 2014, BigDataScience '14.

[21]  Carl E. Rasmussen,et al.  Gaussian processes for machine learning , 2005, Adaptive computation and machine learning.

[22]  Terry Lyons,et al.  Learning from the past, predicting the statistics for the future, learning an evolving system , 2013, 1309.0260.

[23]  Terry Lyons,et al.  Learning stochastic differential equations using RNN with log signature features , 2019, ArXiv.

[24]  École d'été de probabilités de Saint-Flour,et al.  Differential equations driven by rough paths , 2007 .

[25]  Terry Lyons,et al.  Characteristic functions of measures on geometric rough paths , 2013, 1307.3580.