The Signature Kernel

Darrick Lee, Harald Oberhauser
5/8/2023
math.PRcs.LGstat.ML

No official code found. Best implementation: crispitagorico/sigkernel (64 stars, Apache-2.0, last commit 2025-11-04). 1 more below.

Abstract

The signature kernel is a positive definite kernel for sequential data. It inherits theoretical guarantees from stochastic analysis, has efficient algorithms for computation, and shows strong empirical performance. In this short survey paper for a forthcoming Springer handbook, we give an elementary introduction to the signature kernel and highlight these theoretical and computational properties.

View on arXivView PDF

Code Implementations(2)

crispitagorico/sigkernelCommunity100%
Apache-2.0

Differentiable computations for the signature-PDE-kernel on CPU and GPU.

6421Cython, PythonMay 9, 202011 months agoApache-2.0
tgcsaba/KSigCommunity100%
Apache-2.0

A scikit-learn compatible Python package for GPU-accelerated computation of the signature kernel using CuPy.

527Python, Jupyter NotebookMay 10, 20211 years agoApache-2.0

Similar Papers

A User's Guide to $\texttt{KSig}$: GPU-Accelerated Computation of the Signature Kernel

83% match1 repo522025

Random Fourier Signature Features

82% match1 repo522023

Kernel Subspace and Feature Extraction

77% match2023

Notes on Kernel Methods in Machine Learning

76% match2025

Scalable Machine Learning Algorithms using Path Signatures

75% match2025

Log-PDE Methods for Rough Signature Kernels

75% match2024

Cite this paper

@article{lee2023signature,
  title  = {The Signature Kernel},
  author = {Darrick Lee and Harald Oberhauser},
  year   = {2023},
  eprint = {2305.04625},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2305.04625v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2305.04625.svg)](https://codeofpaper.com/paper/2305.04625)

Discussion