Universal approximation theorems for continuous functions of càdlàg paths and Lévy-type signature models

Abstract

We prove two versions of a universal approximation theorem that allow to approximate continuous functions of càdlàg (rough) paths via linear functionals of their time-extended signature, one with respect to the Skorokhod J 1 -topology and the other one with respect to (a rough path version of) the Skorokhod M 1 -topology. Our main motivation to treat this question comes from signature-based models for finance that allow for the inclusion of jumps. Indeed, as an important application, we define a new class of universal signature models based on an augmented Lévy process, which we call Lévy-type signature models. They extend continuous signature models for asset prices as proposed e.g. by in several directions, while still preserving universality and tractability properties. To analyze this, we first show that the signature process of a generic multivariate Lévy process is a polynomial process on the extended tensor algebra and then use this for pricing and hedging approaches within Lévy-type signature models.

A Brief Survey on the Approximation Theory for Sequence Modelling

“…T = R or Z. We note that while sequence modelling is in effect an infinite-dimensional approximation problem, it should be contrasted with generic operator learning problems [9,15,20,47,61,65,75]. Here, a sequence is not a generic function but one with domain being a completely ordered index set.…”