Abstract | ||
---|---|---|
In this paper, we develop an approach to recursively estimate the quadratic risk for matrix recovery problems regularized with spectral functions. Toward this end, in the spirit of the SURE theory, a key step is to compute the (weak) derivative and divergence of a solution with respect to the observations. As such a solution is not available in closed form, but rather through a proximal splitting algorithm, we propose to recursively compute the divergence from the sequence of iterates. A second challenge that we unlocked is the computation of the (weak) derivative of the proximity operator of a spectral function. To show the potential applicability of our approach, we exemplify it on a matrix completion problem to objectively and automatically select the regularization parameter. |
Year | Venue | Keywords |
---|---|---|
2012 | international conference on machine learning | nuclear norm |
Field | DocType | Volume |
Mathematical optimization,Matrix completion,Matrix (mathematics),Matrix function,Quadratic equation,Regularization (mathematics),Operator (computer programming),State-transition matrix,Iterated function,Mathematics | Journal | abs/1205.1482 |
Citations | PageRank | References |
5 | 0.43 | 2 |
Authors | ||
5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Charles-Alban Deledalle | 1 | 387 | 24.00 |
Samuel Vaiter | 2 | 50 | 8.39 |
Gabriel Peyré | 3 | 5 | 0.43 |
Jalal Fadili | 4 | 1184 | 80.08 |
Charles Dossal | 5 | 96 | 8.41 |