Conservative Online Convex Optimization

被引:3
|
作者
de Luca, Martino Bernasconi [1 ]
Vittori, Edoardo [1 ]
Trovo, Francesco [1 ]
Restelli, Marcello [1 ]
机构
[1] Politecn Milan, Dipartimento Elettron Informaz & Bioingn, Piazza Leonardo da Vinci 32, Milan, Italy
来源
MACHINE LEARNING AND KNOWLEDGE DISCOVERY IN DATABASES | 2021年 / 12975卷
关键词
Online learning;
D O I
10.1007/978-3-030-86486-6_2
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Online learning algorithms often have the issue of exhibiting poor performance during the initial stages of the optimization procedure, which in practical applications might dissuade potential users from deploying such solutions. In this paper, we study a novel setting, namely conservative online convex optimization, in which we are optimizing a sequence of convex loss functions under the constraint that we have to perform at least as well as a known default strategy throughout the entire learning process, a.k.a. conservativeness constraint. To address this problem we design a meta-algorithm, namely Conservative Projection (CP), that converts any no-regret algorithm for online convex optimization into one that, at the same time, satisfies the conservativeness constraint and maintains the same regret order. Finally, we run an extensive experimental campaign, comparing and analyzing the performance of our meta-algorithm with that of state-of-the-art algorithms.
引用
收藏
页码:19 / 34
页数:16
相关论文
共 50 条
  • [1] Online Learning and Online Convex Optimization
    Shalev-Shwartz, Shai
    FOUNDATIONS AND TRENDS IN MACHINE LEARNING, 2012, 4 (02): : 107 - 194
  • [2] Boosting for Online Convex Optimization
    Hazan, Elad
    Singh, Karan
    INTERNATIONAL CONFERENCE ON MACHINE LEARNING, VOL 139, 2021, 139
  • [3] Predictive online convex optimization
    Lesage-Landry, Antoine
    Shames, Iman
    Taylor, Joshua A.
    AUTOMATICA, 2020, 113
  • [4] A piecewise conservative method for unconstrained convex optimization
    Scagliotti, A.
    Franzone, P. Colli
    COMPUTATIONAL OPTIMIZATION AND APPLICATIONS, 2022, 81 (01) : 251 - 288
  • [5] A piecewise conservative method for unconstrained convex optimization
    A. Scagliotti
    P. Colli Franzone
    Computational Optimization and Applications, 2022, 81 : 251 - 288
  • [6] An online convex optimization-based framework for convex bilevel optimization
    Shen, Lingqing
    Nam Ho-Nguyen
    Kilinc-Karzan, Fatma
    MATHEMATICAL PROGRAMMING, 2023, 198 (02) : 1519 - 1582
  • [7] An online convex optimization-based framework for convex bilevel optimization
    Lingqing Shen
    Nam Ho-Nguyen
    Fatma Kılınç-Karzan
    Mathematical Programming, 2023, 198 : 1519 - 1582
  • [8] Online Submodular Maximization via Online Convex Optimization
    Salem, Tareq Si
    Ozcan, Gozde
    Nikolaou, Iasonas
    Terzi, Evimaria
    Ioannidis, Stratis
    THIRTY-EIGHTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE, VOL 38 NO 13, 2024, : 15038 - 15046
  • [9] Online Convex Optimization With Binary Constraints
    Lesage-Landry, Antoine
    Taylor, Joshua A.
    Callaway, Duncan S.
    IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 2021, 66 (12) : 6164 - 6170
  • [10] Online convex optimization for cumulative constraints
    Yuan, Jianjun
    Lamperski, Andrew
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 31 (NIPS 2018), 2018, 31