no code implementations • 12 Oct 2021 • Yingshi Chen, Xinli Song, HanYang Dai, Tao Liu, Wuzhi Zhong, Guoyang Wu
The simulation on some large power systems shows that our solver is 2-6 times faster than KLU, which is the state-of-the-art sparse solver for circuit simulation problems.
1 code implementation • 30 Sep 2021 • Yingshi Chen
Our study is the first step to connect these two classical mathematical models: Gaussian Elimination and Markov Decision Process.
no code implementations • 12 Jul 2021 • Yingshi Chen
3) The update of these tokens is a Brownian motion.
no code implementations • 1 Jan 2021 • Yingshi Chen
This CG-FAC method is matrix-free, that is, no need to generate the FIM matrix, also no need to generate the Kronecker factors A and G. We prove that the time and memory complexity of iterative CG-FAC is much less than that of standard K-FAC algorithm.
no code implementations • 27 Oct 2020 • Yingshi Chen
Differentiable forest is actually decision tree based neural turing machine.
1 code implementation • 2 Oct 2020 • Yingshi Chen
So TAB block would learn the importance of each tree and adjust its weight to improve accuracy.
no code implementations • 7 Apr 2020 • Yingshi Chen, Naixing Feng
The deep differentiable forest (DDF) model is introduced to replace the conventional perfectly matched layer (PML) ABC during the computation process of FDTD.
1 code implementation • 29 Feb 2020 • Yingshi Chen
We present a general architecture of deep differentiable forest and its sparse attention mechanism.
1 code implementation • 26 Jan 2020 • Yingshi Chen
In many cases, it only need the data source itself and no extra memory.
1 code implementation • 6 Jan 2020 • Yingshi Chen
Express Wavenet is an improved optical diffractive neural network.
1 code implementation • 23 Dec 2019 • Yingshi Chen, Jinfeng Zhu
Diffractive deep neural network (DNNet) is a novel machine learning framework on the modulation of optical transmission.
no code implementations • 7 Dec 2019 • Yingshi Chen, Jinfeng Zhu
So the the cost of predicted structure is much lower than standard deep CNN.