Xiaoyu Wang – Publication

Preprints

• Xiaoyu Wang and Mikael Johansson.
  On Uniform Boundedness Properties of SGD and its Momentum Variants , arXiv:2201.10245, 2022.

• Xiaoyu Wang and Mikael Johansson.
  Bandwidth-based Step Sizes for Non-Convex Stochastic Optimization, arXiv:2106.02888, 2021.

Journal Papers

• Xiaoyu Wang and Ya-xiang Yuan.
  On the Convergence of Stochastic Gradient Descent with Bandwidth-based Step Size
  Journal of Machine Learning Research, 24(48):1–49, 2023

• Sarit Khirirat, Xiaoyu Wang, Sindri Magnússon, and Mikael Johansson.
  Improved Step-Size Schedules for Proximal Noisy Gradient Methods
  IEEE Transactions on Signal Processing, 2023

• Xiaoyu Wang and Ya-xiang Yuan.
  Stochastic Trust Region Methods with Trust Region Radius Depending on Probabilistic Models.
  Journal of Computational Mathematics, 2022, 40(2): 294-334. DOI: 10.4208/jcm.2012-m2020-0144.

• Xiaoyu Wang, Xiao Wang, and Ya-xiang Yuan.
  Stochastic proximal quasi-Newton methods for non-convex composite optimization.
  Optimization Methods and Software, Volume 34, Issue 5, 2019.

Conference Papers

• Rui Pan, Yuxing Liu, Xiaoyu Wang and Tong Zhang.
  Accelerated Convergence of Stochastic Heavy Ball Method under Anisotropic Gradient Noise
  The 24 International Conference on Learning Representations (ICLR), 2024.

• Xiaoyu Wang, Mikael Johansson and Tong Zhang.
  Generalized Polyak Step Size for First Order Optimization with Momentum.
  Proceedings of the 40th International Conference on Machine Learning (ICML), 2023.

• Renjie Pi, Weizhong Zhang, Yueqi Xie, Jiahui Gao, Xiaoyu Wang, Sunghun Kim, Qifeng Chen.
  Dynafed: Tackling client data heterogeneity with global dynamics.
  Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR), 2023.

• Erik Berglund, Sarit Khirirat, and Xiaoyu Wang.
  Zeroth-order random subspace Newton method.
  accept to ICASSP IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2022.

• Xiaoyu Wang, Sindri Magnússon, and Mikael Johansson.
  On the Convergence of Step Decay Step-Size for Stochastic Optimization.
  35th Conference on Neural Information Processing Systems (NeurIPS 2021), 2021.

• Sarit Khirirat, Xiaoyu Wang, Sindri Magnússon, Mikael Johansson.
  Improved Step-Size Schedules for Noisy Gradient Methods.
  ICASSP 2021-2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2021.