machine learning examples / benchmarks

[martinjrobins]

• dense neural ODE. this blog post by Sebastian Callh is a nice example (https://martinjrobins.github.io/diffsol/primer/weather_neural_ode.html)
• sparse neural ODE. Might be valuable using a sparse neural network for spatial weather forecasting?
• dense/sparse neural ODE with constraints: add algebraic constraint equations
• dense continuous normalising flow (CNF). I'm just looking through the CNF example on the diffrax documentation: https://docs.kidger.site/diffrax/examples/continuous_normalising_flow/
• sparse continuous normalising flow (CNF). I think there would be value in using diffsol to implement a sparse CNF (i.e the network would be sparse so pixels are only connected to their close neighbours) to scale this example up to larger image sizes

Could use burn for this.