Coordinate Descent with Bandit Sampling

被引:0
作者
Salehi, Farnood [1 ]
Thiran, Patrick [1 ]
Celis, L. Elisa [1 ]
机构
[1] Ecole Polytech Fed Lausanne, Sch Comp & Commun Sci, Lausanne, Switzerland
来源
ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 31 (NIPS 2018) | 2018年 / 31卷
关键词
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Coordinate descent methods usually minimize a cost function by updating a random decision variable (corresponding to one coordinate) at a time. Ideally, we would update the decision variable that yields the largest decrease in the cost function. However, finding this coordinate would require checking all of them, which would effectively negate the improvement in computational tractability that coordinate descent is intended to afford. To address this, we propose a new adaptive method for selecting a coordinate. First, we find a lower bound on the amount the cost function decreases when a coordinate is updated. We then use a multi-armed bandit algorithm to learn which coordinates result in the largest lower bound by interleaving this learning with conventional coordinate descent updates except that the coordinate is selected proportionately to the expected decrease. We show that our approach improves the convergence of coordinate descent methods both theoretically and experimentally.
引用
收藏
页数:11
相关论文
共 50 条
[31]   Adam with Bandit Sampling for Deep Learning [J].
Liu, Rui ;
Wu, Tianyi ;
Mozafari, Barzan .
ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS (NEURIPS 2020), 2020, 33
[32]   Asynchronous Parallel Greedy Coordinate Descent [J].
You, Yang ;
Lian, XiangRu ;
Liu, Ji ;
Yu, Hsiang-Fu ;
Dhillon, Inderjit S. ;
Demmel, James ;
Hsieh, Cho-Jui .
ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 29 (NIPS 2016), 2016, 29
[33]   Markov chain block coordinate descent [J].
Sun, Tao ;
Sun, Yuejiao ;
Xu, Yangyang ;
Yin, Wotao .
COMPUTATIONAL OPTIMIZATION AND APPLICATIONS, 2020, 75 (01) :35-61
[34]   Coordinate Descent Full Configuration Interaction [J].
Wang, Zhe ;
Li, Yingzhou ;
Lu, Jianfeng .
JOURNAL OF CHEMICAL THEORY AND COMPUTATION, 2019, 15 (06) :3558-3569
[35]   SparseNet: Coordinate Descent With Nonconvex Penalties [J].
Mazumder, Rahul ;
Friedman, Jerome H. ;
Hastie, Trevor .
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 2011, 106 (495) :1125-1138
[36]   Differentially private block coordinate descent [J].
Riaz, Shazia ;
Ali, Saqib ;
Wang, Guojun ;
Anees, Asad .
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES, 2023, 35 (01) :283-295
[37]   Differentially Private Stochastic Coordinate Descent [J].
Damaskinos, Georgios ;
Mendler-Duenner, Celestine ;
Guerraoui, Rachid ;
Papandreou, Nikolaos ;
Parnell, Thomas .
THIRTY-FIFTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE, THIRTY-THIRD CONFERENCE ON INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE AND THE ELEVENTH SYMPOSIUM ON EDUCATIONAL ADVANCES IN ARTIFICIAL INTELLIGENCE, 2021, 35 :7176-7184
[38]   Accelerating Coordinate Descent in Iterative Reconstruction [J].
Hsieh, Scott S. ;
Hoffman, John M. ;
Noo, Frederic .
MEDICAL IMAGING 2019: PHYSICS OF MEDICAL IMAGING, 2019, 10948
[39]   Stability and Generalization for Randomized Coordinate Descent [J].
Wang, Puyu ;
Wu, Liang ;
Lei, Yunwen .
PROCEEDINGS OF THE THIRTIETH INTERNATIONAL JOINT CONFERENCE ON ARTIFICIAL INTELLIGENCE, IJCAI 2021, 2021, :3104-3110
[40]   Bandit-NAS: Bandit sampling and training method for Neural Architecture Search [J].
Lin, Yiqi ;
Endo, Yuki ;
Lee, Jinho ;
Kamijo, Shunsuke .
NEUROCOMPUTING, 2024, 597