Nonlinear Inverse Problems

Many real forward models are nonlinear: d = G(m) where G is no longer a matrix (seismic travel times, electrical resistivity, chemical kinetics). Superposition fails and there is no closed-form inverse. The standard strategy is to linearize and iterate: solve a sequence of linear inverse problems, each a local approximation, until they converge.

Linearize with the Jacobian

Around a current guess m_k, a first-order Taylor expansion replaces G by its Jacobian J_{ij} = \partial G_i/\partial m_j — the matrix of sensitivities of each datum to each parameter:

G(m_k + \delta m) \approx G(m_k) + J\,\delta m.

Fitting the leftover misfit \Delta d = d - G(m_k) is now a linear least-squares problem for the update \delta m, solved with the very tools from earlier — normal equations plus regularization. Update, recompute the Jacobian, repeat:

m_{k+1} = m_k + (J^{\mathsf T}J + \alpha^2 I)^{-1} J^{\mathsf T}\,\Delta d.

This is the regularized Gauss–Newton iteration (the Levenberg–Marquardt method adapts \alpha as it goes). It is also exactly gradient-based optimization of the misfit — nonlinear inversion is nonlinear least squares.