Zobrazeno 1 - 10
of 1 010
pro vyhledávání: '"ZHANG ShiHua"'
Deep neural networks (DNNs) are vulnerable to small adversarial perturbations of the inputs, posing a significant challenge to their reliability and robustness. Empirical methods such as adversarial training can defend against particular attacks but
Externí odkaz:
http://arxiv.org/abs/2408.00329
Autor:
Gai, Kuo, Zhang, Shihua
In practice, deeper networks tend to be more powerful than shallow ones, but this has not been understood theoretically. In this paper, we find the analytical solution of a three-layer network with a matrix exponential activation function, i.e., $$ f
Externí odkaz:
http://arxiv.org/abs/2407.02540
Publikováno v:
Jixie qiangdu, Vol 42, Pp 551-558 (2020)
It has important theoretical significance and application value for analyzing the fracture failure process of 6061-T6 aluminum alloy to study the relationship between fracture strain and stress triaxiality and establish the fracture failure model. Th
Externí odkaz:
https://doaj.org/article/b8b09c8fed3f44bda16f08258e365cbf
Publikováno v:
Jixie qiangdu, Vol 41, Pp 1436-1444 (2019)
As a parameter characterizes the ability to resist crack propagation,fracture toughness plays an important role in evaluation of pipeline integrity. It is of both scientific and engineering significance to carry out a thorough study on its size effec
Externí odkaz:
https://doaj.org/article/bb9c8508dba54104a88376603b2ec6c0
Neural collapse (NC) is a simple and symmetric phenomenon for deep neural networks (DNNs) at the terminal phase of training, where the last-layer features collapse to their class means and form a simplex equiangular tight frame aligning with the clas
Externí odkaz:
http://arxiv.org/abs/2405.00985
Autor:
Zhang, Rui, Meng, Qi, Zhu, Rongchan, Wang, Yue, Shi, Wenlei, Zhang, Shihua, Ma, Zhi-Ming, Liu, Tie-Yan
In scenarios with limited available data, training the function-to-function neural PDE solver in an unsupervised manner is essential. However, the efficiency and accuracy of existing methods are constrained by the properties of numerical algorithms,
Externí odkaz:
http://arxiv.org/abs/2302.05104
Autor:
Yuan, Jiamiao, Dong, Kangning, Wu, Haixu, Zeng, Xuerui, Liu, Xingyan, Liu, Yan, Dai, Jiapei, Yin, Jichao, Chen, Yongjie, Guo, Yongbo, Luo, Wenhao, Liu, Na, Sun, Yan, Zhang, Shihua, Su, Bing
Publikováno v:
In Cell Genomics 11 December 2024 4(12)
Publikováno v:
In ISPRS Journal of Photogrammetry and Remote Sensing December 2024 218 Part B:466-480
Publikováno v:
In CIRP Journal of Manufacturing Science and Technology November 2024 54:63-74
Autor:
Lai, Pan, Tao, Yiran, Qin, Jun, Xie, Yuanai, Zhang, Shihua, Tang, Shanjiang, Huang, Qirui, Liao, Shengquan
Publikováno v:
In Computer Networks November 2024 253