Search

Conditional p-center problem is one of the classical facility location problems, which aims to find p facilities meeting the given distance condition with q pre-existing facilities. It is worth noting that, with the proliferation of the social network, the effect of social factor is non-negligible. However, the state-of-the-art solutions for the conditional p-center problem fail to deal with the social-aware scenario, and the key challenge is that the existing methods are incompatible with integrating the social condition. In this paper, we first formalize the conditional p-center problem in geo-social networks (GSCpC). To tackle this problem, we develop a homophily-based relaxation algorithm by considering both social constraint and spatial constraint. Specifically, in the first phase, we propose a partition algorithm based on Voronoi diagram to conquer the social constraint. Next, in the second phase, we propose a heuristic algorithm to refine the query results by minimizing the distance. To accelerate the performance, we also propose a category-aware user grouping strategy and a spatial distance based strategy to prune the unpromising results significantly. Experimental results on real-world datasets demonstrate that our approaches are more efficient and effective compared to the benchmarks. [ABSTRACT FROM AUTHOR]

Knowledge graph completion aims to solve the problem of incompleteness and sparsity in knowledge graphs. However, the negative sampling strategy in current completion methods samples entities with equal probability and cannot guarantee the quality of negative samples, causing gradient disappearance during the training process. In addition, existing embedding learning methods ignore the diversity and complexity of the entities and relations in a knowledge graph. Therefore, we design a dynamic semantic sampling and correlation embedding completion framework that includes a negative sampling algorithm based on dynamic semantic similarity and a correlation embedding model. The negative-sampling algorithm can gradually explore high-quality negative samples to participate in model training. The embedding model can enrich embeddings by learning the sequential and correlated information of entities and relations in the knowledge graph. Finally, we conducted experiments on two public datasets, and the experimental results proved the performance of our method. [ABSTRACT FROM AUTHOR]