Search

Traditional model compression techniques are dependent on handcrafted features and require domain experts, with a tradeoff between model size, speed, and accuracy. This study proposes a new approach toward resolving model compression problems. Our approach combines reinforcement-learning-based automated pruning and knowledge distillation to improve the pruning of unimportant network layers and the efficiency of the compression process. We introduce a new state quantity that controls the size of the reward and an attention mechanism that reinforces useful features and attenuates useless features to enhance the effects of other features. The experimental results show that the proposed model is superior to other advanced pruning methods in terms of the computation time and accuracy on CIFAR-100 and ImageNet dataset, where the accuracy is approximately 3% higher than that of similar methods with shorter computation times. [ABSTRACT FROM AUTHOR]

To solve the problem of complex network models with a large number of redundant parameters, a pruning algorithm combined with an attention mechanism is proposed. Firstly, the basic training is performed once, and the network model is then re-trained with the attention mechanism for the baseline. The obtained model is pruned based on channel correlation, and finally a simplified model is obtained via continuous cyclic iteration while the accuracy rate is kept as close as possible to that of the baseline model. The algorithm was experimentally validated on ResNet based on different datasets, and the results showed that the algorithm provided strong adaptability to different datasets and different network structures. For the CIFAR-100 dataset, ResNet50 was pruned to reduce the amount of model parameters by 80.3% and the amount of computation by 69.4%, while maintaining accuracy. For the ImageNet dataset, the ResNet50 parameter volume was compressed by 2.49 times and the computational volume was compressed by 3.01 times. The ResNet101 parameter volume was reduced by 61.2%, and the computational volume was reduced by 68.5%. Compared with the traditional fixed threshold, the model achieves better results in terms of detection accuracy, compression effect, and inference speed. [ABSTRACT FROM AUTHOR]

To further reduce the size of the neural network model and enable the network to be deployed on mobile devices, a novel fusion pruning algorithm based on information entropy stratification is proposed in this paper. Firstly, the method finds similar filters and removes redundant parts by Affinity Propagation Clustering, then secondly further prunes the channels by using information entropy stratification and batch normalization (BN) layer scaling factor, and finally restores the accuracy training by fine-tuning to achieve a reduced network model size without losing network accuracy. Experiments are conducted on the vgg16 and Resnet56 network using the cifar10 dataset. On vgg16, the results show that, compared with the original model, the parametric amount of the algorithm proposed in this paper is reduced by 90.69% and the computation is reduced to 24.46% of the original one. In ResNet56, we achieve a 63.82%-FLOPs reduction by removing 63.53% parameters. The memory occupation and computation speed of the new model are better than the baseline model while maintaining a high network accuracy. Compared with similar algorithms, the algorithm has obvious advantages in the dimensions of computational speed and model size. The pruned model is also deployed to the Internet of Things (IoT) as a target detection system. In addition, experiments show that the proposed model is able to detect targets accurately with low reasoning time and memory. It takes only 252.84 ms on embedded devices, thus matching the limited resources of IoT. [ABSTRACT FROM AUTHOR]

In order to solve the problem of large model computing power consumption, this paper proposes a novel model compression algorithm. Firstly, this paper proposes an interpretable weight allocation method for the loss between a student network (a network model with poor performance), a teacher network (a network model with better performance) and real label. Then, different from the previous simple pruning and fine-tuning, this paper performs knowledge distillation on the pruned model, and quantifies the residual weights of the distilled model. The above operations can further reduce the model size and calculation cost while maintaining the model accuracy. The experimental results show that the weight allocation method proposed in this paper can allocate a relatively appropriate weight to the teacher network and real tags. On the cifar-10 dataset, the pruning method combining knowledge distillation and quantization can reduce the memory size of resnet32 network model from 3726 KB to 1842 KB, and the accuracy can be kept at 93.28%, higher than the original model. Compared with similar pruning algorithms, the model accuracy and operation speed are greatly improved. [ABSTRACT FROM AUTHOR]

Given a source node and a node set, the node-to-set disjoint paths problem is to find disjoint paths from the source node to all nodes from the node set. In this paper, we propose an improved algorithm for the node-to-set disjoint paths problem on bi-rotator graphs. Our result improves the complexity of previous result from to , where n is the dimension of the input graph. [ABSTRACT FROM AUTHOR]

An introduction to the journal is presented in which the editor discusses various reports within the issue on topics including communication systems, networking for pervasive computing environments, and communication algorithms and protocols.

• We prove that a highly scalable data center network (HSDC) is vertex-symmetric. • We investigate the independent spanning trees (IST) problem on HSDC. • We propose an algorithm to construct n ISTs on the n -dimensional HSDC, denoted as H S n. • The algorithm can be parallelized to run in O (n) time using N = 2 n processors. • As a by-product, we obtain the connectivity of H S n to be n. An emerging datacenter network (DCN) with high scalability called HSDC is a server-centric DCN that can help cloud computing in supporting many inherent cloud services. For example, a server-centric DCN can initiate routing for data transmission. This paper investigates the construction of independent spanning trees (ISTs for short), a set of the rooted spanning trees associated with the disjoint-path property, in HSDC. Regarding multiple spanning trees as routing protocol, ISTs have applications in data transmission, e.g., fault-tolerant broadcasting and secure message distribution. We first establish the vertex-symmetry of HSDC. Then, by the structure that n -dimensional HSDC is a compound graph of an n -dimensional hypercube Q n and n -clique K n , we amend the algorithm constructing ISTs for Q n to obtain the algorithm required by HSDC. Unlike most algorithms of recursively constructing tree structures, our algorithm can find every node's parent in each spanning tree directly via an easy computation relied upon only the node address and tree index. Consequently, we can implement the algorithm for constructing n ISTs in O (n N) time, where N = n 2 n is the number of vertices of n -dimensional HSDC; or parallelize the algorithm in O (n) time using N processors. Remarkably, the diameter of the constructed ISTs is about twice the diameter of Q n. [ABSTRACT FROM AUTHOR]

Abstract: Let be a class of graphs. A graph has -width if there are independent sets in such that can be embedded into a graph such that for every edge in which is not an edge in , there exists an such that both endvertices of are in . For the class of block graphs we show that graphs with -width at most 4 are perfect. We also show that -width is NP-complete and show that it is fixed-parameter tractable. For the class of complete graphs, similar results are also obtained. [Copyright &y& Elsevier]

Abstract: Given a class of graphs , a graph is a probe graph of if its vertices can be partitioned into a set of probes and an independent set of nonprobes such that can be embedded into a graph of by adding edges between certain nonprobes. If the partition of the vertices is part of the input, we call a partitioned probe graph of . In this paper we show that there exists a polynomial-time algorithm for the recognition of partitioned probe graphs of comparability graphs. This immediately leads to a polynomial-time algorithm for the recognition of partitioned probe graphs of cocomparability graphs. We then show that a partitioned graph is a partitioned probe permutation graph if and only if is at the same time a partitioned probe graph of comparability and cocomparability graphs. [Copyright &y& Elsevier]