Feature/documentation request: Bayesian neural networks

[DilumAluthge]

It would be great to:

1. Have examples/tutorials of Bayesian neural networks in Soss.jl.
2. Make those examples/tutorials work with SossMLJ.jl.

Prior art:

• Bayesian neural networks in Julia (Turing.jl + Flux.jl): https://turing.ml/dev/tutorials/3-bayesnn/
• Neural nets in MLJ.jl (MLJ.jl + Flux.jl): https://github.com/alan-turing-institute/MLJFlux.jl

[DilumAluthge]

See also: https://github.com/cscherrer/SossMLJ.jl/issues/95

[DilumAluthge]

I'm going to close this issue in favor of https://github.com/cscherrer/SossMLJ.jl/issues/95

The ultimate plan will be to document a full example of a Bayesian neural network that uses SossMLJ; this example will live in SossMLJExamples.jl.