Zobrazeno 1 - 10
of 113
pro vyhledávání: '"Liang, Shiyu"'
Closed-source large language models deliver strong performance but have limited downstream customizability. Semi-open models, combining both closed-source and public layers, were introduced to improve customizability. However, parameters in the close
Externí odkaz:
http://arxiv.org/abs/2410.11182
Autor:
Kang, Huquan, Fu, Luoyi, Funk, Russell J., Wang, Xinbing, Ding, Jiaxin, Liang, Shiyu, Wang, Jianghao, Zhou, Lei, Zhou, Chenghu
The past few centuries have witnessed a dramatic growth in scientific and technological knowledge. However, the nature of that growth - whether exponential or otherwise - remains controversial, perhaps partly due to the lack of quantitative character
Externí odkaz:
http://arxiv.org/abs/2409.08349
Autor:
Liang, Shiyu, Wang, Ziyuan, Huang, Zhenghua, Wei, Hengyuan, Fu, Hui, Xiong, Ming, Xia, Lidong
Loops are fundamental structures in the magnetized atmosphere of the sun. Their physical properties are crucial for understanding the nature of the solar atmosphere. Transition region loops are relatively dynamic and their physical properties have no
Externí odkaz:
http://arxiv.org/abs/2404.09805
Autor:
Lu, Bin, Ma, Tingyan, Gan, Xiaoying, Wang, Xinbing, Zhu, Yunqiang, Zhou, Chenghu, Liang, Shiyu
Graph Neural Networks (GNNs) are widely deployed in vast fields, but they often struggle to maintain accurate representations as graphs evolve. We theoretically establish a lower bound, proving that under mild conditions, representation distortion in
Externí odkaz:
http://arxiv.org/abs/2404.04969
Autor:
Wang, Xinbing, Fu, Luoyi, Gan, Xiaoying, Wen, Ying, Zheng, Guanjie, Ding, Jiaxin, Xiang, Liyao, Ye, Nanyang, Jin, Meng, Liang, Shiyu, Lu, Bin, Wang, Haiwen, Xu, Yi, Deng, Cheng, Zhang, Shao, Kang, Huquan, Wang, Xingli, Li, Qi, Guo, Zhixin, Qi, Jiexing, Liu, Pan, Ren, Yuyang, Wu, Lyuwen, Yang, Jungang, Zhou, Jianping, Zhou, Chenghu
The exponential growth of scientific literature requires effective management and extraction of valuable insights. While existing scientific search engines excel at delivering search results based on relational databases, they often neglect the analy
Externí odkaz:
http://arxiv.org/abs/2403.02576
Graph Neural Network (GNN) has demonstrated extraordinary performance in classifying graph properties. However, due to the selection bias of training and testing data (e.g., training on small graphs and testing on large graphs, or training on dense g
Externí odkaz:
http://arxiv.org/abs/2308.08344
Publikováno v:
In Algal Research December 2024 84
Recent theoretical works on over-parameterized neural nets have focused on two aspects: optimization and generalization. Many existing works that study optimization and generalization together are based on neural tangent kernel and require a very lar
Externí odkaz:
http://arxiv.org/abs/2104.11895
One of the major concerns for neural network training is that the non-convexity of the associated loss functions may cause bad landscape. The recent success of neural networks suggests that their loss landscape is not too bad, but what specific resul
Externí odkaz:
http://arxiv.org/abs/2007.01429
Traditional landscape analysis of deep neural networks aims to show that no sub-optimal local minima exist in some appropriate sense. From this, one may be tempted to conclude that descent algorithms which escape saddle points will reach a good local
Externí odkaz:
http://arxiv.org/abs/1912.13472