Séminaire ISTerre

---

Autoregressive shaping regularization

jeudi 13 octobre 2022 - 10h30

Sergey Fomel (+ Ludovic Métivier) - Jackson School of Geosciences, University of Texas at Austin
--- 
Regularization is a method of imposing additional constraints on the estimated model in the process of solving inverse problems. Shaping regularization implements such a constrain by applying a shaping operator, which maps the estimated model to the space of admissible models. In this presentation, I connect shaping with covariance estimation by regularized non-stationary autoregression (adaptive prediction-error filters) and show how to implement it efficiently using fast explicit diffusion. This approach finds applications in a wide variety of geophysical inverse problems.

---

Equipe organisatrice : Grands Séminaires ISTerre

Amphithéâtre Killian, Maison des Géosciences, 38400 Saint Martin d'Hères

Informations de visio :

https://univ-grenoble-alpes-fr.zoom.us/j/95687850018?pwd=eGFqeUtnN1RSUTJRdzlsdnppTXVXdz09