ϕℓ=σ(ωℓ(wx+b)) where σ is an arbitrary activation function.
h(x;θ(x)) For example we have the hypernetwork:
ϕℓ(x;θ)[wθ,bθ]=wθx+bθ=h(z;θ) We have the additive transformation
ϕℓ(x;θ)sθ=wℓx+bℓ+sθ=h(z;θ) We have the affine transformation
ϕℓ(x;θ)[sθ,aθ]=(wℓx+bℓ)sθ+aθ=h(z;θ) Example: 2D Sea Surface Height¶
In this example, we will train the same SIREN model from before but we will add a modulation extension.