In stochastic analysis, a rough path is a generalization of the classical notion of a smooth path. It extends calculus and differential equation theory to handle irregular signals—paths that are too rough for traditional analysis, such as a Wiener process. This makes it possible to define and solve controlled differential equations of the form d y t = f ( y t ) d x t , y 0 = a {\displaystyle \mathrm {d} y_{t}=f(y_{t})\mathrm {d} x_{t},\quad y_{0}=a} even when the driving path x t {\displaystyle x_{t}} lacks classical differentiability. The theory was introduced in the 1990s by Terry Lyons. Rough path theory captures how nonlinear systems interact with highly oscillatory or noisy input. It builds on the integration theory of L. C. Young, the geometric algebra of Kuo-Tsai Chen, and the Lipschitz function theory of Hassler Whitney, while remaining compatible with key ideas in stochastic calculus. The theory also extends Itô's theory of stochastic differential equations far beyond the semimartingale setting. Its definitions and uniform estimates form a robust framework that can recover classical results—such as the Wong–Zakai theorem, the Stroock–Varadhan support theorem, and the construction of stochastic flows—without relying on probabilistic properties like martingales or predictability. A central concept in the theory is the Signature of a path: a noncommutative transform that encodes the path as a sequence of iterated integrals. Formally, it is a homomorphism from the monoid of paths (under concatenation) into the group-like elements of a tensor algebra. The Signature is faithful—it uniquely characterizes paths up to certain negligible modifications—making it a powerful tool for representing and comparing paths. These iterated integrals play a role similar to monomials in a Taylor expansion: they provide a coordinate system that captures the essential features of a path. Just as Taylor’s theorem allows a smooth function to be approximated locally by polynomials, the terms of the Signature offer a structured, hierarchical summary of a path’s behavior. This enriched representation forms the basis for defining a rough path and enables analysis without directly examining its fine-scale structure. The theory has widespread applications across mathematics and applied fields. Notably, Martin Hairer used rough path techniques to help construct a solution theory for the KPZ equation, and later developed the more general theory of regularity structures, for which he was awarded the Fields Medal in 2014.
Motivation Rough path theory aims to make sense of the controlled differential equation
d Y t i = ∑ j = 1 d V j i ( Y t ) d X t j . {\displaystyle \mathrm {d} Y_{t}^{i}=\sum _{j=1}^{d}V_{j}^{i}(Y_{t})\,\mathrm {d} X_{t}^{j}.}
where the control, the continuous path X t {\displaystyle X_{t}} taking values in a Banach space, need not be differentiable nor of bounded variation. A prevalent example of the controlled path X t {\displaystyle X_{t}} is the sample path of a Wiener process. In this case, the aforementioned controlled differential equation can be interpreted as a stochastic differential equation and integration against " d X t j {\displaystyle \mathrm {d} X_{t}^{j}} " can be defined in the sense of Itô. However, Itô's calculus is defined in the sense of L 2 {\displaystyle L^{2}} and is in particular not a pathwise definition. Rough paths give an almost sure pathwise definition of stochastic differential equations. The rough path notion of solution is well-posed in the sense that if X ( n ) t {\displaystyle X(n)_{t}} is a sequence of smooth paths converging to X t {\displaystyle X_{t}} in the p {\displaystyle p} -variation metric (described below), and
d Y ( n ) t i = ∑ j = 1 d V j i ( Y t ) d X ( n ) t j ; {\displaystyle \mathrm {d} Y(n)_{t}^{i}=\sum _{j=1}^{d}V_{j}^{i}(Y_{t})\,\mathrm {d} X(n)_{t}^{j};}
… excerpt ends here. Continue reading the full article.
