Abstract:Signature-based techniques give mathematical insight into the interactions between complex streams of evolving data. These insights can be quite naturally translated into numerical approaches to understanding streamed data, and perhaps because of their mathematical precision, have proved useful in analysing streamed data in situations where the data is irregular, and not stationary, and the dimension of the data and the sample sizes are both moderate. Understanding streamed multi-modal data is exponential: a word in $n$ letters from an alphabet of size $d$ can be any one of $d^n$ messages. Signatures remove the exponential amount of noise that arises from sampling irregularity, but an exponential amount of information still remain. This survey aims to stay in the domain where that exponential scaling can be managed directly. Scalability issues are an important challenge in many problems but would require another survey article and further ideas. This survey describes a range of contexts where the data sets are small enough to remove the possibility of massive machine learning, and the existence of small sets of context free and principled features can be used effectively. The mathematical nature of the tools can make their use intimidating to non-mathematicians. The examples presented in this article are intended to bridge this communication gap and provide tractable working examples drawn from the machine learning context. Notebooks are available online for several of these examples. This survey builds on the earlier paper of Ilya Chevryev and Andrey Kormilitzin which had broadly similar aims at an earlier point in the development of this machinery. This article illustrates how the theoretical insights offered by signatures are simply realised in the analysis of application data in a way that is largely agnostic to the data type.

On expected signatures and signature cumulants in semimartingale models

Unified Signature Cumulants and Generalized Magnus Expansions

Signature Methods in Machine Learning

Generalized Euler-Maclaurin formula and Signatures

A multiplicative surface signature through its Magnus expansion

A representation for the Expected Signature of Brownian motion up to the first exit time of the planar unit disc

Signature Methods in Stochastic Portfolio Theory

Gaussian processes based data augmentation and expected signature for time series classification

A PDE approach for solving the characteristic function of the generalised signature process

Almost sure approximations and laws of iterated logarithm for signatures

On the representation property for 1D general diffusion semimartingales

Embedding and learning with signatures

Generalized iterated-sums signatures

Expected Signature on a Riemannian Manifold and Its Geometric Implications

Concentration and Exact Convergence Rates for Expected Brownian Signatures

On the signature of an image

A Primer on the Signature Method in Machine Learning

Primal and dual optimal stopping with signatures

Martingale representation in progressive enlargement by the reference filtration of a semimartingale: a note on the multidimensional case

Fractional signature: a generalisation of the signature inspired by fractional calculus

Asymptotic properties of realized power variations and related functionals of semimartingales