no code implementations • 10 Mar 2024 • Xunpeng Huang, Hanze Dong, Difan Zou, Tong Zhang
Along this line, Freund et al. (2022) suggest that the modified Langevin algorithm with prior diffusion is able to converge dimension independently for strongly log-concave target distributions.
no code implementations • 12 Jan 2024 • Xunpeng Huang, Difan Zou, Hanze Dong, Yian Ma, Tong Zhang
Specifically, DMC follows the reverse SDE of a diffusion process that transforms the target distribution to the standard Gaussian, utilizing a non-parametric score estimation.
no code implementations • 5 Jul 2023 • Xunpeng Huang, Hanze Dong, Yifan Hao, Yi-An Ma, Tong Zhang
We propose a Monte Carlo sampler from the reverse diffusion process.
no code implementations • 19 May 2022 • Jingwei Zhang, Xunpeng Huang
We consider optimizing two-layer neural networks in the mean-field regime where the learning dynamics of network weights can be approximated by the evolution in the space of probability measures over the weight parameters associated with the neurons.
no code implementations • 1 Jan 2021 • Xunpeng Huang, Vicky Jiaqi Zhang, Hao Zhou, Lei LI
Adaptive gradient methods have been shown to outperform SGD in many tasks of training neural networks.
1 code implementation • 12 Jun 2020 • Xunpeng Huang, Runxin Xu, Hao Zhou, Zhe Wang, Zhengyang Liu, Lei LI
Due to its simplicity and outstanding ability to generalize, stochastic gradient descent (SGD) is still the most widely used optimization method despite its slow convergence.
no code implementations • 12 Jun 2020 • Xunpeng Huang, Hao Zhou, Runxin Xu, Zhe Wang, Lei LI
Adaptive gradient methods have attracted much attention of machine learning communities due to the high efficiency.
no code implementations • 10 Feb 2020 • Xunpeng Huang, Xianfeng Liang, Zhengyang Liu, Yitan Li, Linyun Yu, Yue Yu, Lei LI
SPAN computes the inverse of the Hessian matrix via low-rank approximation and stochastic Hessian-vector products.
no code implementations • 25 Sep 2019 • Xunpeng Huang, Zhengyang Liu, Zhe Wang, Yue Yu, Lei LI
To the best of our knowledge, Acutum is the first adaptive gradient method without second moments.
2 code implementations • 11 Nov 2017 • Junliang Guo, Linli Xu, Xunpeng Huang, Enhong Chen
In this paper, we take a matrix factorization perspective of network embedding, and incorporate structure, content and label information of the network simultaneously.