Inhalt des Dokuments
zur Navigation
Publikationsliste
| Zitatschlüssel | CsSchOpSa16 |
|---|---|
| Autor | Botond Cseke and David Schnoerr and Manfred Opper and Guido Sanguinetti |
| Seiten | 494002 |
| Jahr | 2016 |
| DOI | http://dx.doi.org/10.1088/1751-8113/49/49/494002 |
| Journal | Journal of Physics A: Mathematical and Theoretical |
| Jahrgang | 49 |
| Nummer | 49 |
| Verlag | IOPscience |
| Zusammenfassung | We consider the inverse problem of reconstructing the posterior measure over the trajectories of a diffusion process from discrete time observations and continuous time constraints. We cast the problem in a Bayesian framework and derive approximations to the posterior distributions of single time marginals using variational approximate inference, giving rise to an expectation propagation type algorithm. For non-linear diffusion processes, this is achieved by leveraging moment closure approximations. We then show how the approximation can be extended to a wide class of discrete-state Markov jump processes by making use of the chemical Langevin equation. Our empirical results show that the proposed method is computationally efficient and provides good approximations for these classes of inverse problems. |