Software

I’m an author / developer / major contributor for several pieces of software, mostly to do with differential equations in the PyTorch ecosystem. See also my GitHub page.

torchcde (Author)

Controlled differential equation (CDE) solvers with differentiation and GPU support. Backpropagation through the solver or via the adjoint method is supported; the latter allows for improved memory efficiency.

In particular this allows for building Neural Controlled Differential Equation models, which are state-of-the-art models for (arbitrarily irregular!) time series. Neural CDEs can be thought of as a “continuous time RNN”.

torchsde (Developer)

Stochastic differential equation (SDE) solvers with differentiation and GPU support. Backpropagation through the solver or via the adjoint method is supported; the latter allows for improved memory efficiency.

torchdiffeq (Major contributor)

Ordinary differential equation (ODE) solvers with differentation and GPU support. Backpropagation through the solver or via the adjoint method is supported; the latter allows for improved memory efficiency.

Signatory (Author)

Differentiable computations of the signature and logsignature transforms, on both CPU and GPU.