I just wrote this first version when I was learning Haskell and so I went for simplicity and since theoretically we’re guaranteed that it’s enough, I just stopped there. However, we’re preparing (for a long time now) a new version. But let me warn you that I’m quite busy, so is the other guy working on HNN so we’re a bit short on time and don’t know at all when it’ll be available. So let me know if you’re interested!

is there possible to build neural network with multiple layers in HNN? From its sources it seams, that it can have only two tiers:

check :: [[Neuron]] -> Bool
check nss = let l = length nss in l > 1 && l [[Neuron]]
nn nss | check nss = nss
| otherwise = error “Invalid nn”

Why it can’t take here any number of layers?

best regards,
Robert