Tighter Generalization Bounds for Matrix Completion Via Factorization Into Constrained Matrices
Tighter Generalization Bounds for Matrix Completion Via Factorization Into Constrained Matrices
复制标题
DOI：
10.1587/transinf.2017edp7339
复制
发表时间：
2018-08
期刊：
影响因子：
--
通讯作者：
Ken-ichiro Moridomi;Kohei Hatano;Eiji Takimoto
中科院分区：
文献类型：
--
作者：
Ken-ichiro Moridomi;Kohei Hatano;Eiji Takimoto
我们证明了具有范数约束的低秩矩阵类的泛化误差界。与仅使用秩或相关量的已知边界相比，通过考虑额外的l1和l1约束，我们的边界更加紧密。此外，我们还证明了类的Rademacher复杂度的边界是最优的。
SUMMARY We prove generalization error bounds of classes of low-rank matrices with some norm constraints for collaborative filtering tasks. Our bounds are tighter, compared to known bounds using rank or the re-lated quantity only, by taking the additional L 1 and L 1 constraints into account. Also, we show that our bounds on the Rademacher complexity of the classes are optimal.