Code for "Neural Controlled Differential Equations for Irregular Time Series" (Neurips 2020 Spotlight)

Differentiable computations of the signature and logsignature transforms, on both CPU and GPU. (ICLR 2021)

Code for: "Neural Rough Differential Equations for Long Time Series", (ICML 2021)

Code for "Deep Signature Transforms" (NeurIPS 2019)

SigKAN: Signature-Weighted Kolmogorov-Arnold Networks for Time Series

Methodology for anomaly detection on multivariate streams using path signatures and the variance norm.

Compute rough path signature of a drawn path

An elementary level assignment for Visualization of Rough Paths using Python.

A library for numeric calculation of everything rough paths and RDEs.