His research interests include graph partitioning and clustering, parallel algorithms and combinatorial optimization in the context of big data. His graph partitioning algorithms – KaHIP – have been able to i
Partitioning through projections: Strong SDP bounds for large graph partition problemsGraph partition problemsSemidefinite programmingCutting planesDykstra?s projection algorithmAugmented Lagrangian methodsThe graph partition problem (GPP) aims at clustering the vertex set of a graph into a fixed number of ...
Graph partitioning is a key problem to enable efficient solving of a wide range of computational tasks and querying over large-scale graph data, such as computing node centralities using iterative computations, and personalized recommendations. In this work, we introduce a unifying...
(10) 聚类(clustering) 一个超图的clustering C=\{C_1,...,C_l\} 是其顶点集合的分区。 与k-way分区相对,簇(clusters)的数量不会预先提供。并且对于 C_i 的实际大小不施加平衡限制。 (11) k-way hypergraph edge partitioning (k路超图边分区) 一个H 的k-way超图边分区是将其超边集合分区到 k 个...
Graph clustering is a form ofgraph miningthat is useful in a number ofpractical applications including marketing, customer segmentation, congestiondetection, facility location, and XML data integration (Lee, Hsu, Yang, &Yang,2002). The graph clustering problems are typically defined into twocategories...
YAN, JTYan (" Two-way balance-tolerant partitioning based on fuzzy graph clustering for hierarchical design og VLSI systems ", IEEE, Conference Proceedings of the 1995 IEEE Fourteenth Annual International Phoenix Conference on Computers and Communications, 28 Mar....
Currently, research in the realm of fast graph clustering is experiencing an explosive growth; however, a comprehensive and systematic survey is noticeably absent. To bridge this gap, our focus will primarily be on reviewing methods from two perspectives: single-view and multi-view. The main struc...
In this paper we develop a novel parallel spectral partitioning method that takes advantage of an efficient implementation of a preconditioned eigenvalue solver and a k-means algorithm on the GPU. We showcase the performance of our novel scheme against standard spectral techniques. Also, we use it...
Many state-of-the-art scalability approaches tackle this challenge by sampling neighborhoods for mini-batch training, graph clustering and partitioning, or by using simplified GNN models. These approaches have been implemented in PyG, and can benefit from the above GNN layers, operators and models....
This coefficient matrix reflects the relationships between the original data samples and anchor data samples well, so utilizing the coefficient matrix effectively for clustering is a research direction. To handle this, Dhillon et al. has already studied the bipartite spectral graph partitioning algorithm...