Non-adversarial training of Neural SDEs with signature kernel scores

Z Issa, B Horvath, M Lemercier… - Advances in Neural …, 2023 - proceedings.neurips.cc
Neural SDEs are continuous-time generative models for sequential data. State-of-the-art
performance for irregular time series generation has been previously obtained by training …

Neural signature kernels as infinite-width-depth-limits of controlled resnets

NM Cirone, M Lemercier… - … Conference on Machine …, 2023 - proceedings.mlr.press
Motivated by the paradigm of reservoir computing, we consider randomly initialized
controlled ResNets defined as Euler-discretizations of neural controlled differential …

Efficient and accurate gradients for neural sdes

P Kidger, J Foster, XC Li… - Advances in Neural …, 2021 - proceedings.neurips.cc
Neural SDEs combine many of the best qualities of both RNNs and SDEs, and as such are a
natural choice for modelling many types of temporal dynamics. They offer memory efficiency …

Koopman kernel regression

P Bevanda, M Beier, A Lederer… - Advances in …, 2023 - proceedings.neurips.cc
Many machine learning approaches for decision making, such as reinforcement learning,
rely on simulators or predictive models to forecast the time-evolution of quantities of interest …

Lecture notes on rough paths and applications to machine learning

T Cass, C Salvi - ar** via distribution regression: a higher rank signature approach
B Horvath, M Lemercier, C Liu, T Lyons… - ar** the
law of a stochastic process to a scalar target. The learning procedure based on the notion of …

Random fourier signature features

C Toth, H Oberhauser, Z Szabo - arxiv preprint arxiv:2311.12214, 2023 - arxiv.org
Tensor algebras give rise to one of the most powerful measures of similarity for sequences
of arbitrary length called the signature kernel accompanied with attractive theoretical …