no code implementations • 2 Dec 2022 • Yang Shi, Guannan Liang, Young-joo Chung
Training samples in RS can be highly biased toward popular businesses with sufficient sales and can decrease advertising performance for small businesses.
no code implementations • 12 Apr 2021 • Tan Zhu, Guannan Liang, Chunjiang Zhu, Haining Li, Jinbo Bi
In this work, we formulate the SCB that uses a DNN reward function as a non-convex stochastic optimization problem, and design a stage-wise stochastic gradient descent algorithm to optimize the problem and determine the action policy.
no code implementations • 7 Mar 2021 • Guannan Liang, Qianqian Tong, Chunjiang Zhu, Jinbo Bi
Stochastically controlled stochastic gradient (SCSG) methods have been proved to converge efficiently to first-order stationary points which, however, can be saddle points in nonconvex optimization.
no code implementations • 31 Dec 2020 • Qianqian Tong, Guannan Liang, Tan Zhu, Jinbo Bi
Nonconvex sparse learning plays an essential role in many areas, such as signal processing and deep network compression.
no code implementations • 14 Sep 2020 • Guannan Liang, Qianqian Tong, Jiahao Ding, Miao Pan, Jinbo Bi
Sparse learning is a very important tool for mining useful information and patterns from high dimensional data.
no code implementations • 14 Sep 2020 • Qianqian Tong, Guannan Liang, Jinbo Bi
Federated learning allows loads of edge computing devices to collaboratively learn a global model without data sharing.
no code implementations • 11 Aug 2020 • Jiahao Ding, Jingyi Wang, Guannan Liang, Jinbo Bi, Miao Pan
In PP-ADMM, each agent approximately solves a perturbed optimization problem that is formulated from its local private data in an iteration, and then perturbs the approximate solution with Gaussian noise to provide the DP guarantee.
2 code implementations • 2 Aug 2019 • Qianqian Tong, Guannan Liang, Jinbo Bi
Theoretically, we provide a new way to analyze the convergence of AGMs and prove that the convergence rate of \textsc{Adam} also depends on its hyper-parameter $\epsilon$, which has been overlooked previously.
no code implementations • NeurIPS 2016 • Jin Lu, Guannan Liang, Jiangwen Sun, Jinbo Bi
We prove that when the side features can span the latent feature space of the matrix to be recovered, the number of observed entries needed for an exact recovery is $O(\log N)$ where $N$ is the size of the matrix.