Corpus ID: 67855865

Logarithmic Regret for parameter-free Online Logistic Regression

@article{Vilmarest2019LogarithmicRF,
  title={Logarithmic Regret for parameter-free Online Logistic Regression},
  author={Joseph De Vilmarest and O. Wintenberger},
  journal={ArXiv},
  year={2019},
  volume={abs/1902.09803}
}
We consider online optimization procedures in the context of logistic regression, focusing on the Extended Kalman Filter (EKF). We introduce a second-order algorithm close to the EKF, named Semi-Online Step (SOS), for which we prove a O(log(n)) regret in the adversarial setting, paving the way to similar results for the EKF. This regret bound on SOS is the first for such parameter-free algorithm in the adversarial logistic regression. We prove for the EKF in constant dynamics a O(log(n)) regret… Expand

References

SHOWING 1-10 OF 14 REFERENCES
Logarithmic regret algorithms for online convex optimization
  • 833
  • Highly Influential
  • PDF
Non-strongly-convex smooth stochastic approximation with convergence rate O(1/n)
  • 284
  • Highly Influential
  • PDF
Optimal non-asymptotic bound of the Ruppert-Polyak averaging without strong convexity
  • 15
  • PDF
Kullback-Leibler aggregation and misspecified generalized linear models
  • 61
  • PDF
Introduction to Online Convex Optimization
  • Elad Hazan
  • Computer Science, Mathematics
  • Found. Trends Optim.
  • 2016
  • 778
  • Highly Influential
  • PDF
Online natural gradient as a Kalman filter
  • 34
  • PDF
Posterior Mode Estimation by Extended Kalman Filtering for Multivariate Dynamic Generalized Linear Models
  • 161
  • Highly Influential
New Results in Linear Filtering and Prediction Theory
  • 5,687
  • PDF
...
1
2
...