Neural controlled differential equations for irregular time series

P Kidger, J Morrill, J Foster… - Advances in Neural …, 2020 - proceedings.neurips.cc
Neural ordinary differential equations are an attractive option for modelling temporal
dynamics. However, a fundamental issue is that the solution to an ordinary differential …

Deep learning for human activity recognition on 3d human skeleton: survey and comparative study

HC Nguyen, TH Nguyen, R Scherer, VH Le - Sensors, 2023 - mdpi.com
Human activity recognition (HAR) is an important research problem in computer vision. This
problem is widely applied to building applications in human–machine interactions …

Neural rough differential equations for long time series

J Morrill, C Salvi, P Kidger… - … Conference on Machine …, 2021 - proceedings.mlr.press
Neural controlled differential equations (CDEs) are the continuous-time analogue of
recurrent neural networks, as Neural ODEs are to residual networks, and offer a memory …

Sig-Wasserstein GANs for time series generation

H Ni, L Szpruch, M Sabate-Vidales, B ** the path signature methodology and its application to landmark-based human action recognition
W Yang, T Lyons, H Ni, C Schmid, L ** - Stochastic Analysis, Filtering …, 2022 - Springer
Landmark-based human action recognition in videos is a challenging task in computer
vision. One key step is to design a generic approach that generates discriminative features …

Monitoring and prediction of landslide-related deformation based on the GCN-LSTM algorithm and SAR imagery

MA Khalili, L Guerriero, M Pouralizadeh, D Calcaterra… - Natural Hazards, 2023 - Springer
A key component of disaster management and infrastructure organization is predicting
cumulative deformations caused by landslides. One of the critical points in predicting …

Framing RNN as a kernel method: A neural ODE approach

A Fermanian, P Marion, JP Vert… - Advances in Neural …, 2021 - proceedings.neurips.cc
Building on the interpretation of a recurrent neural network (RNN) as a continuous-time
neural differential equation, we show, under appropriate conditions, that the solution of a …

A generalised signature method for multivariate time series feature extraction

J Morrill, A Fermanian, P Kidger, T Lyons - arxiv preprint arxiv:2006.00873, 2020 - arxiv.org
The'signature method'refers to a collection of feature extraction techniques for multivariate
time series, derived from the theory of controlled differential equations. There is a great deal …

Summarization of videos with the signature transform

J de Curtò, I de Zarzà, G Roig, CT Calafate - Electronics, 2023 - mdpi.com
This manuscript presents a new benchmark for assessing the quality of visual summaries
without the need for human annotators. It is based on the Signature Transform, specifically …