Weighted low rank approximations with provable guarantees

110 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Recovery guarantee of weighted low-rank approximation via alternating minimization (2016, background): “On the theoretical side, the only result we know of is [Razenshteyn et al., 2016], who provide a ﬁxed-parameter tractability result when additionally the weight matrix is low-rank.”
Recovery Guarantees for One-hidden-layer Neural Networks (2017, methodology): “However, by making some assumptions, tensor decomposition method becomes efﬁcient (Anandkumar et al., 2014; Wang et al., 2015; Wang & Anandkumar, 2016; Song et al., 2016).”
Towards Fast Computation of Certified Robustness for ReLU Networks (2018, methodology): “By making a stronger assumption of ETH, we can have the following stronger result which excludes all 2o(n c) time algorithms, where c > 0 is some fixed constant: Corollary A.18.”
Efficient Low-rank Multimodal Fusion With Modality-Specific Factors (2018, methodology): “Razenshteyn et al. (2016) proposes a mod-iﬁed weighted version of low-rank approximation, and Koch and Lubich (2010) applies the method towards temporally dependent data to obtain low-rank approximations.”
Low rank approximation with entrywise l1-norm error (2017, methodology): “This algorithm also leverages a technique called “guessing a sketch” from [77].”