Scalable gradients for stochastic differential equations

X Li, TKL Wong, RTQ Chen… - … Conference on Artificial …, 2020 - proceedings.mlr.press
The adjoint sensitivity method scalably computes gradients of solutions to ordinary
differential equations. We generalize this method to stochastic differential equations …

Ode2vae: Deep generative second order odes with bayesian neural networks

C Yildiz, M Heinonen… - Advances in Neural …, 2019 - proceedings.neurips.cc
Abstract We present Ordinary Differential Equation Variational Auto-Encoder (ODE2VAE), a
latent second order ODE model for high-dimensional sequential data. Leveraging the …

Macroscopic traffic flow modeling with physics regularized Gaussian process: A new insight into machine learning applications in transportation

Y Yuan, Z Zhang, XT Yang, S Zhe - Transportation Research Part B …, 2021 - Elsevier
Despite the wide implementation of machine learning (ML) technique in traffic flow modeling
recently, those data-driven approaches often fall short of accuracy in the cases with a small …

Continuous-time model-based reinforcement learning

C Yildiz, M Heinonen… - … Conference on Machine …, 2021 - proceedings.mlr.press
Abstract Model-based reinforcement learning (MBRL) approaches rely on discrete-time state
transition models whereas physical systems and the vast majority of control tasks operate in …

Scalable gradients and variational inference for stochastic differential equations

X Li, TKL Wong, RTQ Chen… - … on Advances in …, 2020 - proceedings.mlr.press
We derive reverse-mode (or adjoint) automatic differentiation for solutions of stochastic
differential equations (SDEs), allowing time-efficient and constant-memory computation of …

Learning interacting dynamical systems with latent gaussian process odes

Ç Yıldız, M Kandemir… - Advances in Neural …, 2022 - proceedings.neurips.cc
We study uncertainty-aware modeling of continuous-time dynamics of interacting objects.
We introduce a new model that decomposes independent dynamics of single objects …

[HTML][HTML] Continual learning from demonstration of robotics skills

S Auddy, J Hollenstein, M Saveriano… - Robotics and …, 2023 - Elsevier
Methods for teaching motion skills to robots focus on training for a single skill at a time.
Robots capable of learning from demonstration can considerably benefit from the added …

FAIR: Fair collaborative active learning with individual rationality for scientific discovery

X Xu, Z Wu, A Verma, CS Foo… - … Conference on Artificial …, 2023 - proceedings.mlr.press
Scientific discovery aims to find new patterns and test specific hypotheses by analysing
large-scale experimental data. However, various practical limitations (eg, high experimental …

Learning efficient and robust ordinary differential equations via invertible neural networks

W Zhi, T Lai, L Ott, EV Bonilla… - … Conference on Machine …, 2022 - proceedings.mlr.press
Advances in differentiable numerical integrators have enabled the use of gradient descent
techniques to learn ordinary differential equations (ODEs), where a flexible function …

Neural graphical modelling in continuous-time: consistency guarantees and algorithms

A Bellot, K Branson, M van der Schaar - arXiv preprint arXiv:2105.02522, 2021 - arxiv.org
The discovery of structure from time series data is a key problem in fields of study working
with complex systems. Most identifiability results and learning algorithms assume the …