Distributed and privacy-preserving federated learning (FL) has been associated with edge computing systems for developing intelligent IoT applications. However, collecting data individually in each FL node may result ...
详细信息
Distributed and privacy-preserving federated learning (FL) has been associated with edge computing systems for developing intelligent IoT applications. However, collecting data individually in each FL node may result in non-independent and identically distributed (non-IID) training data, which can significantly impair FL performance. To address this, we propose a model transfer approach that allows a FL client to train its model on more datasets, collectively forming an IID virtual dataset. In the proposed approach, FL clients are classified as transfer-demand or aggregation-inclined based on their local model's experienced data label distributions. Transfer-demand clients transfer their models to helper clients that can support their lacking labels, while aggregation-inclined clients have enough data labels in their training models, and thus, they participate in the model aggregation. However, two or more transfer-demand clients may contend with the same helper client. To resolve the contention, we apply the minimum-cost maximum-matching (MCMM) framework and integer linear programming to find the optimal solution. To minimize model transmission costs among FL clients, we use a Steiner tree-based solution to dynamically allocate a parameter server that aggregates the local models from clients. Finally, we perform extensive simulation experiments with different problems, and our proposed approach significantly outperforms baseline methods, in the case of MNIST achieves 99.02% accuracy and reduces the communication cost within the range of 20% to 66% and in the case of CIFAR-10 dataset, enhances the accuracy at least 24% and communication cost in the range of 29.48% to 63.09% in comparison with non-IID baseline approaches.
Spectral matching (SM) is an efficient and effective greedy algorithm for solving the graph matching problem in feature correspondence in computer vision and graphics. However, the classic SM algorithm cannot extract ...
详细信息
Spectral matching (SM) is an efficient and effective greedy algorithm for solving the graph matching problem in feature correspondence in computer vision and graphics. However, the classic SM algorithm cannot extract correspondences well when the affinity matrix is sparse and reducible (i.e. its corresponding graph is not connected). This case often happens when the geometric deformations consist of transformations with local inconsistency. The authors analyse this problem and show how the original SM could fail in this scenario. Then, the authors propose a revised two-step pipeline to tackle this issue: (1) decompose the mutually inconsistent local deformations into several consistent transformations which can be solved by individual SM;(2) filter out incorrect correspondences through an automatic thresholding. The authors perform experiments to demonstrate that this modification can effectively handle the coarse correspondence computation in shape or image registration where the global transformation consists of multiple inconsistent local transformations.
Given a set of edge pairs in a complete bipartite graph, we want to find a bipartite matching that includes the maximum number of those edge pairs. While the problem has many applications to wireless localization and ...
详细信息
Given a set of edge pairs in a complete bipartite graph, we want to find a bipartite matching that includes the maximum number of those edge pairs. While the problem has many applications to wireless localization and computer vision, to the best of our knowledge, there is no theoretical work for the problem. In this work, unless P = N P, we show that there is no constant approximation for the problem. Then, we consider two special cases of the problem. Suppose that k denotes the maximum number of input edge pairs such that a particular node can be in. Inspired by experimental results, the first case is for when k is not large. While there is a simple polynomial-time algorithm for the problem when k is one, we show that the problem is NP-hard when k is greater than one. We also devise an efficient O(k)-approximation algorithm for the problem. For the second case, every pair of nodes in the same partition of the input bipartite graph are labeled with one of chi colors. We want to match, between the two partitions, a pair of nodes to a pair of nodes with the same color. Denote n as the number of nodes, we give an O (root chi(n))-approximation algorithm for this case. (C) 2021 Elsevier B.V. All rights reserved.
problems related to graphmatching and isomorphisms are very important both from a theoretical and practical perspective, with applications ranging from image and video analysis to biological and biomedical problems. ...
详细信息
problems related to graphmatching and isomorphisms are very important both from a theoretical and practical perspective, with applications ranging from image and video analysis to biological and biomedical problems. The graph matching problem is challenging from a computational point of view, and therefore different relaxations are commonly used. Although common relaxations techniques tend to work well for matching perfectly isomorphic graphs, it is not yet fully understood under which conditions the relaxed problem is guaranteed to obtain the correct answer. In this paper, we prove that the graph matching problem and its most common convex relaxation, where the matching domain of permutation matrices is substituted with its convex hull of doubly-stochastic matrices, are equivalent for a certain class of graphs, such equivalence being based on spectral properties of the corresponding adjacency matrices. We also derive results about the automorphism group of a graph, and provide fundamental spectral properties of the adjacency matrix.
暂无评论