I’m trying to construct minimal neural networks for a set of functions. the functions are complete – I’m trying to replicate them completely in NN, with minimal resources (non-zero weights).
(Obviously the term “minimal” is very loose, there will be heavy constraints on the contents/structure of the network, but im fine with that)
I saw a few algorithms that do this, but the best algorithm i saw (though it is very old – 1989) is the “Regular partitioning” (link to paper)
- Is there a better algorithm for generating minimal networks? surely there is an algorithm for this that’s been suggested within the last 30 years
- If there isn’t an obvious candidate (and even if there is) – is there an implementation in code for the RP algorithm? I couldnt find anything with a quick search, but I’m not sure where to look further.
In general, where should I look when I want to find code implementations of algorithms? github?