arxiv_cs (![[info]](http://l-stat.livejournal.com/img/userinfo.gif){kind=small} arxiv_cs) wrote,@ 2008-12-11 18:21:00 |
|
Convex Sparse Matrix Factorizations. (arXiv:0812.1869v1 [cs.LG])
Convex Sparse Matrix Factorizations. (arXiv:0812.1869v1 [cs.LG])
read more at cs updates on arXiv.org
Convex Sparse Matrix Factorizations. (arXiv:0812.1869v1 [cs.LG])
We present a convex formulation of dictionary learning for sparse signal decomposition. Convexity is obtained by replacing the usual explicit upper bound on the dictionary size by a convex rank-reducing term similar to the trace norm. In particular, our formulation introduces an explicit trade-off between size and sparsity of the decomposition of rectangular matrices. Using a large set of synthetic examples, we compare the estimation abilities of the convex and non-convex approaches, showing that while the convex formulation has a single local minimum, this may lead in some cases to performance which is inferior to the local minima of the non-convex formulation.
read more at cs updates on arXiv.org
