In this paper we present an efficient algorithm for computing a sparse null space basis for a full row rank matrix. We first apply the ideas of the Markowitz’s pivot selection criterion to a rank reducing algorithm to propose an efficient algorithm for computing sparse null space bases of full row rank matrices. We then describe how we can use the Dulmage-Mendelsohn decomposition to make the re...

Collaborative filtering (CF) recommender systems are very popular and successful in commercial application fields. However, robustness analysis research has shown that conventional memory-based recommender systems are very susceptible to malicious profile-injection attacks. A number of attack models have been proposed and studied and recent work has suggested that model-based CF algorithms have...

Tensor rank and low-rank tensor decompositions have many applications in learning and complexity theory. Most known algorithms use unfoldings of tensors and can only handle rank up to nbp/2c for a p-th order tensor in Rnp . Previously no efficient algorithm can decompose 3rd order tensors when the rank is super-linear in the dimension. Using ideas from sum-of-squares hierarchy, we give the firs...

A novel algorithm is proposed for CANDECOMP/PARAFAC tensor decomposition to exploit best rank-1 tensor approximation. Different from the existing algorithms, our algorithm updates rank-1 tensors simultaneously in-parallel. In order to achieve this, we develop new all-at-once algorithms for best rank-1 tensor approximation based on the Levenberg-Marquardt method and the rotational update. We sho...

k-nearest neighbour (kNN) collaborative filtering (CF), the widely successful algorithm supporting recommender systems, attempts to relieve the problem of information overload by generating predicted ratings for items users have not expressed their opinions about; to do so, each predicted rating is computed based on ratings given by like-minded individuals. Like-mindedness, or similarity-based ...

This paper investigates error and erasure decoding methods for codes with maximum rank distance. These codes can be used for correcting column and row errors and erasures in an ( ) array. Such errors occur e.g. in magnetic tape recording or in memory chip arrays. For maximum rank distance codes (Rank-Codes), there exists a decoding algorithm similar to the Peterson-Gorenstein-Zierler technique ...

We propose a new learning to rank algorithm, named Weighted Margin-Rank Batch loss (WMRB), to extend the popular Weighted Approximate-Rank Pairwise loss (WARP). WMRB uses a new rank estimator and an efficient batch training algorithm. The approach allows more accurate item rank approximation and explicit utilization of parallel computation to accelerate training. In three item recommendation ta...

Conventional collaborative filtering(CF) recommendation applies the user-based centralized architecture. This architecture has some problems of sparsity and scalability, in addition to not fit the current popular P2P architecture. Therefore, this paper proposes a distributed model to implement the CF algorithm by maintaining the user’s record information distributedly in each nodes throughout t...

Citation recommendation plays an important role in the context of scholarly big data, where finding relevant papers has become more difficult because information overload. Applying traditional collaborative filtering (CF) to citation is challenging due cold start problem and lack paper ratings. To address these challenges, this article, we propose a with network representation learning framewor...

AIMS Although contact force (CF)-guided circumferential pulmonary vein isolation (CPVI) for paroxysmal atrial fibrillation (PAF) is useful, AF recurrence at long-term follow-up still remains to be resolved. The purpose of this study was to assess safety and efficacy of CF-guided CPVI and to compare residual conduction gaps during CPVI and long-term outcome between the conventional (non-CF-guide...