Joint Tensor Factorization and Outlying Slab Suppression With Applications
Joint Tensor Factorization and Outlying Slab Suppression With Applications
We consider factoring low-rank tensors in the presence of outlying slabs. This problem is important in practice, because data collected in many real-world applications, such as speech, fluorescence, and some social network data, fit this paradigm. Prior work tackles this problem by iteratively selecting a fixed number of slabs and …