← 1 The adjoint method - Stanford Computer Science