Search

The purpose is to deal with the problems of excessive pruning of important convolutional layers caused by equalscale pruning, retain many redundant parameters and large loss of accuracy. This paper integrated sensitivity analysis on the basis of FPGM pruning strategy for network pruning. The algorithm used precision feedback to analyze the importance of each convolutional layer, control the single-layer pruning ratio, analyze the impact of different pruning ratios in each layer on accuracy loss, and obtained the sensitivity of each convolutional layer. It could be combined with FPGM to analyze the importance of the convolution kernel and cut the unimportant convolution kernel according to the pruning ratio of the sensitivity to complete the neural network pruning. The experimental results show that the accuracy of this method only drops by l. 56% and 0. 11 % when the pruning rate on MobileNet-v 1 and ResNet50 is 50%. This method has a higher pruning rate and lower calculation amount on ResNet50 under the same accuracy loss. [ABSTRACT FROM AUTHOR]

With the improvement of space geodetic observation accuracy, the geodetic inversion model is more precise and detailed, and higher requirements are also put forward for the inversion algorithm. Focusing on the inversion of fault parameters in the spherical layered model, the defects of genetic algorithm is ana- lyzed and a new combinational inversion algorithm based on iterative least squares algorithm and genetic al- gorithm is proposed. The co-seismic deformation inversion mode due to seismic fault has set up considering the earth layering. The effectiveness of the combinational algorithm has been verified using the simulated data. In the inversion of genetic algorithm, it is often difficult to obtain good results of fault depth and slip distance.. The main reason is the strong correlation between these parameters and the secondary reason is the low sensitivity of depth and slip. The combinational algorithm proposed in this paper has achieved ideal results on faults of various dip angles and has practicality. [ABSTRACT FROM AUTHOR]

Focused on dynamic scheduling of single-processor real-time system, this paper defined and analyzed the sensitivity and influence of task attributes including arrival time, execution time, deadline and laxity, then it proposed a scheduling algorithm PTBM based on priority table, which made that a task with small deadline and large laxity had higher priority. Compared with PTBM with EDF, LLF and PTD, simulated results verify CMRN of the sensitivity and influence, and show PTBM outperforming on scheduling success ratio. The proposed algorithm can be applied to dynamically schedule real-time tasks in realtime systems. [ABSTRACT FROM AUTHOR]