Sequential prediction under log-loss with side information

被引：0
|
作者：
Bhatt, Alankrita [1 ]
Kim, Young-Han [1 ]
机构：
[1] Univ Calif San Diego, Dept Elect & Comp Engn, La Jolla, CA 92093 USA
来源：
ALGORITHMIC LEARNING THEORY, VOL 132 | 2021年 / 132卷
关键词：
DATA-COMPRESSION; UNIVERSAL;
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：
The problem of online prediction with sequential side information under logarithmic loss is studied, and general upper and lower bounds on the minimax regret incurred by the predictor is established. The upper bounds on the minimax regret are obtained by constructing and analyzing a probability assignment based on mixture probability assignments in universal compression, and the lower bounds are obtained by way of a redundancy-capacity theorem. A tight characterization of the regret is provided in some special settings.
引用
收藏
页数：5
相关论文
共 50 条
  • [1] Sequential Normalized Maximum Likelihood in Log-loss Prediction
    Kotlowski, Wojciech
    Grunwald, Peter
    [J]. 2012 IEEE INFORMATION THEORY WORKSHOP (ITW), 2012, : 547 - 551
  • [2] Universal Batch Learning with Log-Loss
    Fogel, Yaniv
    Feder, Meir
    [J]. 2018 IEEE INTERNATIONAL SYMPOSIUM ON INFORMATION THEORY (ISIT), 2018, : 21 - 25
  • [3] On the Problem of On-line Learning with Log-Loss
    Fogel, Yaniv
    Feder, Meir
    [J]. 2017 IEEE INTERNATIONAL SYMPOSIUM ON INFORMATION THEORY (ISIT), 2017,
  • [4] Constructing Multiclass Classifiers using Binary Classifiers Under Log-Loss
    Ben-Yishai, Assaf
    Ordentlich, Or
    [J]. 2021 IEEE INTERNATIONAL SYMPOSIUM ON INFORMATION THEORY (ISIT), 2021, : 2435 - 2440
  • [5] A Multiple Description CEO Problem with Log-Loss Distortion
    Pichler, Georg
    Piantanida, Pablo
    Matz, Gerald
    [J]. 2017 IEEE INTERNATIONAL SYMPOSIUM ON INFORMATION THEORY (ISIT), 2017, : 111 - 115
  • [6] Asymptotic log-loss of prequential maximum likelihood codes
    Grünwald, P
    de Rooij, S
    [J]. LEARNING THEORY, PROCEEDINGS, 2005, 3559 : 652 - 667
  • [7] Binary CEO Problem under Log-Loss with BSC Test-Channel Model
    Nangir, Mahdi
    Asvadi, Reza
    Ahmadian-Attari, Mahmoud
    Chen, Jun
    [J]. 2018 29TH BIENNIAL SYMPOSIUM ON COMMUNICATIONS (BSC), 2018,
  • [8] Label Inference Attacks from Log-loss Scores
    Aggarwal, Abhinav
    Kasiviswanathan, Shiva Prasad
    Xu, Zekun
    Feyisetan, Oluwaseyi
    Teissier, Nathanael
    [J]. INTERNATIONAL CONFERENCE ON MACHINE LEARNING, VOL 139, 2021, 139
  • [9] Successive Wyner-Ziv Coding for the Binary CEO Problem under Log-Loss
    Nangir, Mahdi
    Asvadi, Reza
    Ahmadian-Attari, Mahmoud
    Chen, Jun
    [J]. 2018 29TH BIENNIAL SYMPOSIUM ON COMMUNICATIONS (BSC), 2018,
  • [10] Regret Bounds for Log-Loss via Bayesian Algorithms
    Wu, Changlong
    Heidari, Mohsen
    Grama, Ananth
    Szpankowski, Wojciech
    [J]. IEEE TRANSACTIONS ON INFORMATION THEORY, 2023, 69 (09) : 5971 - 5989
12345