Zobrazeno 1 - 10
of 2 146
pro vyhledávání: '"Yang, En"'
Although it is traditionally believed that lossy image compression, such as JPEG compression, has a negative impact on the performance of deep neural networks (DNNs), it is shown by recent works that well-crafted JPEG compression can actually improve
Externí odkaz:
http://arxiv.org/abs/2410.07081
Autor:
Zheng, Kaixiang, Yang, En-Hui
As a technique to bridge logit matching and probability distribution matching, temperature scaling plays a pivotal role in knowledge distillation (KD). Conventionally, temperature scaling is applied to both teacher's logits and student's logits in KD
Externí odkaz:
http://arxiv.org/abs/2402.11148
Publikováno v:
International Conference on Learning Representations 2024 (ICLR)
It is believed that in knowledge distillation (KD), the role of the teacher is to provide an estimate for the unknown Bayes conditional probability distribution (BCPD) to be used in the student training process. Conventionally, this estimate is obtai
Externí odkaz:
http://arxiv.org/abs/2401.08732
Autor:
Hamidi, Shayan Mohajer, Yang, En-Hui
Federated learning (FL) is a promising technology via which some edge devices/clients collaboratively train a machine learning model orchestrated by a server. Learning an unfair model is known as a critical problem in federated learning, where the tr
Externí odkaz:
http://arxiv.org/abs/2401.04993
The concepts of conditional mutual information (CMI) and normalized conditional mutual information (NCMI) are introduced to measure the concentration and separation performance of a classification deep neural network (DNN) in the output probability d
Externí odkaz:
http://arxiv.org/abs/2309.09123
Publikováno v:
Zhongguo shipin weisheng zazhi, Vol 36, Iss 7, Pp 878-885 (2024)
Deoxynivalenol （DON） is a toxic secondary metabolite produced by Fusarium， which mainly infects wheat， corn and other grains. It not only causes huge economic losses to the agricultural industry， but also has a potential threat to human and
Externí odkaz:
https://doaj.org/article/0090a43a0e894966bdfc638d76e04e86
With limited storage/bandwidth resources, input images to Computer Vision (CV) applications that use Deep Neural Networks (DNNs) are often encoded with JPEG that is tailored to Human Vision (HV). This paper presents Deep Selector-JPEG, an adaptive JP
Externí odkaz:
http://arxiv.org/abs/2302.09560
Autor:
Xu, Ruoqing1 (AUTHOR), Yang, En1 (AUTHOR), Liang, Hsin1 (AUTHOR), Luo, Shenying1 (AUTHOR), Liu, Yunhan1 (AUTHOR), Khoong, Yimin1 (AUTHOR), Li, Haizhou1 (AUTHOR), Huang, Xin1 (AUTHOR) huangxindoctor@yahoo.com, Zhao, Yixuan1 (AUTHOR) zhaoyixuan@sjtu.edu.cn, Zan, Tao1 (AUTHOR) zantao@sjtu.edu.cn
Publikováno v:
Clinical & Translational Medicine. Sep2024, Vol. 14 Issue 9, p1-22. 22p.
Autor:
Jia, Tianyu, Yang, En-Yu, Hsiao, Yu-Shun, Cruz, Jonathan, Brooks, David, Wei, Gu-Yeon, Reddi, Vijay Janapa
Autonomous machines (e.g., vehicles, mobile robots, drones) require sophisticated 3D mapping to perceive the dynamic environment. However, maintaining a real-time 3D map is expensive both in terms of compute and memory requirements, especially for re
Externí odkaz:
http://arxiv.org/abs/2205.03325
Publikováno v:
In Heliyon 15 October 2024 10(19)