Search

Your search keyword '"Hu, Xing"' showing total 1,865 results
Start Over Author "Hu, Xing" Publication Type Academic Journals
1,865 results on '"Hu, Xing"'

13. Dipterocarpoidae genomics reveal their demography and adaptations to Asian rainforests

Author

Wang, Rong, Liu, Chao-Nan, Segar, Simon T., Jiang, Yu-Ting, Zhang, Kai-Jian, Jiang, Kai, Wang, Gang, Cai, Jing, Chen, Lu-Fan, Chen, Shan, Cheng, Jing, Compton, Stephen G., Deng, Jun-Yin, Ding, Yuan-Yuan, Du, Fang K., Hu, Xiao-Di, Hu, Xing-Hua, Kang, Ling, Li, Dong-Hai, Lu, Ling, Li, Yuan-Yuan, Tang, Liang, Tong, Xin, Wang, Zheng-Shi, Xu, Wei-Wei, Yang, Yang, Zang, Run-Guo, Zu, Zhuo-Xin, Zhang, Yuan-Ye, and Chen, Xiao-Yong

Published
2024
Full Text
View/download PDF

18. Comprehensive SNN Compression Using ADMM Optimization and Activity Regularization

Author

Deng, Lei, Wu, Yujie, Hu, Yifan, Liang, Ling, Li, Guoqi, Hu, Xing, Ding, Yufei, Li, Peng, and Xie, Yuan

Subjects
Affordable and Clean Energy, Neurons, Computational modeling, Quantization, Optimization, Encoding, Task analysis, Synapses, Activity regularization, alternating direction method of multiplier, connection pruning, spiking neural network (SNN) compression, weight quantization, Artificial Intelligence & Image Processing
Abstract

As well known, the huge memory and compute costs of both artificial neural networks (ANNs) and spiking neural networks (SNNs) greatly hinder their deployment on edge devices with high efficiency. Model compression has been proposed as a promising technique to improve the running efficiency via parameter and operation reduction, whereas this technique is mainly practiced in ANNs rather than SNNs. It is interesting to answer how much an SNN model can be compressed without compromising its functionality, where two challenges should be addressed: 1) the accuracy of SNNs is usually sensitive to model compression, which requires an accurate compression methodology and 2) the computation of SNNs is event-driven rather than static, which produces an extra compression dimension on dynamic spikes. To this end, we realize a comprehensive SNN compression through three steps. First, we formulate the connection pruning and weight quantization as a constrained optimization problem. Second, we combine spatiotemporal backpropagation (STBP) and alternating direction method of multipliers (ADMMs) to solve the problem with minimum accuracy loss. Third, we further propose activity regularization to reduce the spike events for fewer active operations. These methods can be applied in either a single way for moderate compression or a joint way for aggressive compression. We define several quantitative metrics to evaluate the compression performance for SNNs. Our methodology is validated in pattern recognition tasks over MNIST, N-MNIST, CIFAR10, and CIFAR100 datasets, where extensive comparisons, analyses, and insights are provided. To the best of our knowledge, this is the first work that studies SNN compression in a comprehensive manner by exploiting all compressible components and achieves better results.

Published
2023

19. Exploring Adversarial Attack in Spiking Neural Networks With Spike-Compatible Gradient

Author

Liang, Ling, Hu, Xing, Deng, Lei, Wu, Yujie, Li, Guoqi, Ding, Yufei, Li, Peng, and Xie, Yuan

Subjects
Bioengineering, Neurosciences, Spatiotemporal phenomena, Computational modeling, Perturbation methods, Biological neural networks, Backpropagation, Unsupervised learning, Training, Adversarial attack, backpropagation through time, neuromorphic computing, spike-compatible gradient, spiking neural networks, Artificial Intelligence & Image Processing
Abstract

Spiking neural network (SNN) is broadly deployed in neuromorphic devices to emulate brain function. In this context, SNN security becomes important while lacking in-depth investigation. To this end, we target the adversarial attack against SNNs and identify several challenges distinct from the artificial neural network (ANN) attack: 1) current adversarial attack is mainly based on gradient information that presents in a spatiotemporal pattern in SNNs, hard to obtain with conventional backpropagation algorithms; 2) the continuous gradient of the input is incompatible with the binary spiking input during gradient accumulation, hindering the generation of spike-based adversarial examples; and 3) the input gradient can be all-zeros (i.e., vanishing) sometimes due to the zero-dominant derivative of the firing function. Recently, backpropagation through time (BPTT)-inspired learning algorithms are widely introduced into SNNs to improve the performance, which brings the possibility to attack the models accurately given spatiotemporal gradient maps. We propose two approaches to address the above challenges of gradient-input incompatibility and gradient vanishing. Specifically, we design a gradient-to-spike (G2S) converter to convert continuous gradients to ternary ones compatible with spike inputs. Then, we design a restricted spike flipper (RSF) to construct ternary gradients that can randomly flip the spike inputs with a controllable turnover rate, when meeting all-zero gradients. Putting these methods together, we build an adversarial attack methodology for SNNs. Moreover, we analyze the influence of the training loss function and the firing threshold of the penultimate layer on the attack effectiveness. Extensive experiments are conducted to validate our solution. Besides the quantitative analysis of the influence factors, we also compare SNNs and ANNs against adversarial attacks under different attack methods. This work can help reveal what happens in SNN attacks and might stimulate more research on the security of SNN models and neuromorphic devices.

Published
2023

36. Epidemiological and clinical characteristics of functional movement disorder based on movement disorders clinic

Author

WANG Jin⁃tao, MA Sha, WANG Han, BA Mao⁃wen, CHEN Xian⁃wen, HU Xing⁃yue, LIU Yi⁃ming, CHENG Ou⁃mei, WANG Yu⁃kai, ZHOU Fan, XU Shun⁃liang, SHEN Yue⁃fei, CHEN Wei, DOU Rong⁃hua, LIU Xue⁃dong, HAN Yan⁃qing, WANG Hua⁃long, DING Yan, PAN Xiao⁃ling, HUANG Wei, CHEN Lei, LU Hong, XU Gang, and WANG Gang

Subjects
conversion disorder, motor disorders, epidemiology, Neurology. Diseases of the nervous system, RC346-429
Abstract

Objective To preliminarily summarize the epidemiological and clinical characteristics of functional movement disorder (FMD) based on movement disorders clinic in China. Methods and Results A total of 593 patients with movement disorders newly diagnosed in movement disorders clinic of 22 clinical medical centers from August to September in 2023 were included, including 37 patients (6.24%) with FMD and 556 patients (93.76%) with organic movement disorder. Comparison of the epidemiological characteristics of the 2 groups, the proportions of corona virus disease (COVID⁃19) infection (χ2 = 4.217, P = 0.040) and psychopsychological symptoms (χ2 = 18.694, P = 0.000) in the FMD group were higher than those in the organic movement disorder group, while age (t = 3.757, P = 0.000), age of onset (t = 3.720, P = 0.000) and proportion of hypertension (χ2 = 4.736, P = 0.030) were lower than those in the organic movement disorder group. Comparison of the clinical characteristics of the 2 groups, the FMD group had higher proportions of tremor (χ2 = 3.955, P = 0.047) and myoclonus (Fisher′s exact probability: P = 0.011) than the organic movement disorder group, and the proportions of parkinsonism (χ2 = 8.491, P = 0.004) and gait disorder (χ2 = 5.028, P = 0.025) were lower than those in the organic movement disorder group. Conclusions The age and age of onset, as well as the proportions of COVID⁃19 infection, hypertension, psychopsychological symptoms, tremor, myoclonus, parkinsonism and gait disorder with FMD were different from those with organic movement disorder in China.

Published
2023
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results