$$ \large \begin{aligned} &\text{model} : \theta \times \xi \rightarrow (\chi \times \phi) \rightarrow \psi \\ &\text{loss} : \psi \times \psi \rightarrow \ell \\ &\text{optimization} : (\theta \times \xi \times \chi \rightarrow \psi) \rightarrow (\psi \times \psi \rightarrow \ell) \rightarrow \mathcal{D} \rightarrow \theta \times \lambda \times \mathcal{L} \\ &\text{inference} : (\theta \times \kappa \rightarrow \psi) \rightarrow \theta \rightarrow (\chi \times \kappa \times \zeta) \rightarrow \psi \end{aligned} $$