no code implementations • 11 Dec 2023 • Yueyao Yu, Yin Zhang
Since its introduction in 2017, Transformer has emerged as the leading neural network architecture, catalyzing revolutionary advancements in many AI disciplines.
no code implementations • 17 May 2022 • Xinyu Chen, Renjie Li, Yueyao Yu, Yuanwen Shen, Wenye Li, Zhaoyu Zhang, Yin Zhang
In this work, we propose the first-ever Transformer model (POViT) to efficiently design and simulate semiconductor photonic devices with multiple objectives.
no code implementations • 29 Sep 2021 • Yueyao Yu, Yin Zhang
We introduce a notion of variability to view such issues under the setting of a fixed number of parameters which is, in general, a dominant cost-factor.
1 code implementation • 8 Jun 2021 • Yueyao Yu, Yin Zhang
To enhance resource efficiency and model deployability of neural networks, we propose a neural-layer architecture based on Householder weighting and absolute-value activating, called Householder-absolute neural layer or simply Han-layer.
no code implementations • 19 May 2021 • Yueyao Yu, Yin Zhang
Despite the tremendous successes of deep neural networks (DNNs) in various applications, many fundamental aspects of deep learning remain incompletely understood, including DNN trainability.
no code implementations • 1 Jan 2021 • Yueyao Yu, Jie Wang, Wenye Li, Yin Zhang
The stochastic gradient descent (SGD) method, first proposed in 1950's, has been the foundation for deep-neural-network (DNN) training with numerous enhancements including adding a momentum or adaptively selecting learning rates, or using both strategies and more.
no code implementations • 18 Feb 2019 • Yueyao Yu, Pengfei Yu, Wenye Li
Deep learning models are vulnerable to adversarial examples, which poses an indisputable threat to their applications.