Skip to content

Analysis of Linear Functions

Let \(\psi : \IR^K \ra \IR^M, \phi: \IR^M \ra \IR^N, a: \IR^N \ra \IR\) be continuously differentiable functions. We have the following notions of derivatives:

  • Jacobi Matrix - \(D_p \phi \in \IR[M,N]\) is a MxN matrix of partial derivatives: \(D_p \phi[m,n] = (\del_n \phi^m) (p)\).
  • Nabla Operators. We define \(\nabla a (p) = (D_p a)^* \in \IR[N,1]=\IR^N\) as transpose of the Jacobi Matrix.

The following computational rules hold true:

Chain Rule

\[ D_p(\phi \circ \psi) = ( D_{\psi(p)} \phi ) \cdot (D_p \phi), \qquad \nabla_p(a \circ \phi) = (D_p \phi)^* \cdot \nabla_{\phi(p)} a \]

Linear Maps

  • If \(\phi=A\) is a linear map, represented by a matrix \(A \in \IR[M,N]\), then

    \[ D_p A = A. \]
  • For \(v \in \IR^N\) and \(f(p)=\<v,p\>\) we have:

    \[ D_p f = f = v^* = \< v,\_\>, \qquad \nabla f (p) = v. \]

Quadratic Forms

  • Let \(f(p) = \half \| p\|^2 = \half \< p, p \>\), then

    \[ D_p f = \< p, \_\> = p^*, \qquad \nabla f (p) = p. \]
  • Let \(f(p) = \half <G p, p>\), then

    \[ D_p f = \half \< (G+G^*)p, \_ \>, \qquad \nabla f(p) = \half (G+G^*)p \]

    If \(G=G^*\), this simplifies to:

    \[ D_p f = \<G p,\_ \> = p^* G^*, \qquad \nabla f (p) = G p. \]
  • Let \(f(p) = \half \| A p \|^2\), then

    \[ D_p f = \<A^* Ap,\_ \> = p^* A^*A, \qquad \nabla f (p) = A^*A p. \]
  • Let \(f(p) = \half \| A^* p - t \|^2\), then

    \[ D_p f = \< A^* Ap - A^* t,\_\> \qquad \nabla f(p) = A^* A p - A^* t. \]