Francesco Martinuzzi

Novel minimal reservoir computing

1

**What kind of problems is it mostly used for?** Short time series modeling, showing improved results over NGRC (see #108) **Other implementations to know about** No open implementations available at...

https://arxiv.org/abs/2010.07103

https://arxiv.org/abs/2211.05262v1

https://journals.aps.org/pre/abstract/10.1103/PhysRevE.104.014205 https://journals.aps.org/prresearch/abstract/10.1103/PhysRevResearch.3.013090

Using `@example`s in the docs at the moment doesn't allow for GPU example, the Mackey Glass forecasting script is stil available [here](https://github.com/SciML/ReservoirComputing.jl/blob/master/docs/src/esn_tutorials/mackeyglass_basic.md). Going forward this example needs a way to...

Changes on the forcing of same type as input data on the internals have broken the ability to input boolean valued arrays to the RECA models

the need for the one dimensional and two dimensional types is to dispatch over the `CellularAutomaton` function, but it's restricting in creating new CAs, since they need to be built...

[WIP] Types extension

2

This PR is meant to extend the package to return output with matching typing of the given input data. Specifically I am adding the flexibility to all functions to handle...

Hardcoding of `Float64` in loss

1

Is there a motivation for [hardcoding](https://github.com/JuliaAI/MLJLinearModels.jl/blob/dev/src/loss-penalty/generic.jl#L21) the scaling in the loss penalties as `Float64`? If not, would a more generic definition allow for multiple types of regression outputs (Float64, 32,...