Regularization with zero-inflation

In response to the sentence
> software for some of the more exotic models used by ecologists (e.g. zero-inflated models) may not be readily available

I note that you can ridge using glmmTMB, so that software **is** available (see also https://github.com/glmmTMB/glmmTMB/issues/827, though whether L2 should be used for variable selection is another discussion entirely). However, I am sure that there are other model types where good software for regularization is missing.

It might be good to cite Wood (2017) somewhere in this area, since it includes extensive discussion on methods for selecting tuning parameters in penalized regression (implemented in mgcv).