Amortized efficiency of list update and paging rules
Communications of the ACM
COLT '90 Proceedings of the third annual workshop on Computational learning theory
The weighted majority algorithm
Information and Computation
Predicting nearly as well as the best pruning of a decision tree
COLT '95 Proceedings of the eighth annual conference on Computational learning theory
A dynamic disk spin-down technique for mobile computing
MobiCom '96 Proceedings of the 2nd annual international conference on Mobile computing and networking
Journal of the ACM (JACM)
A decision-theoretic generalization of on-line learning and an application to boosting
Journal of Computer and System Sciences - Special issue: 26th annual ACM symposium on the theory of computing & STOC'94, May 23–25, 1994, and second annual Europe an conference on computational learning theory (EuroCOLT'95), March 13–15, 1995
On-line learning and the metrical task system problem
COLT '97 Proceedings of the tenth annual conference on Computational learning theory
Online computation and competitive analysis
Online computation and competitive analysis
On-line Algorithms in Machine Learning
Developments from a June 1996 seminar on Online algorithms: the state of the art
Semi-supervised approach to rapid and reliable labeling of large data sets
Proceedings of the 14th ACM SIGKDD international conference on Knowledge discovery and data mining
Learning to Trade with Incremental Support Vector Regression Experts
HAIS '08 Proceedings of the 3rd international workshop on Hybrid Artificial Intelligence Systems
A randomized online learning algorithm for better variance control
COLT'06 Proceedings of the 19th annual conference on Learning Theory
Taming traffic dynamics: Analysis and improvements
Computer Communications
Hi-index | 0.00 |
This work is concerned with online learning from expert advice. Extensive work on this problem generated numerous "expert advice algorithms" whose total loss is provably bounded above in terms of the loss incurred by the best expert in hindsight. Such algorithms were devised for various problem variants corresponding to various loss functions. For some loss functions, such as the square, Hellinger and entropy losses, optimal algorithms are known. However, for two of the most widely used loss functions, namely the 0/1 and absolute loss, there are still gaps between the known lower and upper bounds.In this paper we present two new expert advice algorithms and prove for them the best known 0/1 and absolute loss bounds. Given an expert advice algorithm ALG, the goal is to form an upper bound on the regret L ALG -- L* of ALG, where L ALG is the loss of ALG and L* is the loss of the best expert in hindsight. Typically, regret bounds of a "canonical form" C · $$\sqrt {L^ * \ln N} $$ are sought where N is the number of experts and C is a constant. So far, the best known constant for the absolute loss function is C = 2.83, which is achieved by the recent IAWM algorithm of Auer et al. (2002). For the 0/1 loss function no bounds of this canonical form are known and the best known regret bound is $$L_{ALG} - L* \leqslant L* + C_1 \ln N + C_2 \sqrt {L*\ln N + \frac{e}{4}\ln ^2 N} $$ , where C 1 = e 驴 2 and C 2 = 2 $$\sqrt e $$ . This bound is achieved by a "P-norm" algorithm of Gentile and Littlestone (1999). Our first algorithm is a randomized extension of the "guess and double" algorithm of Cesa-Bianchi et al. (1997). While the guess and double algorithm achieves a canonical regret bound with C = 3.32, the expected regret of our randomized algorithm is canonically bounded with C = 2.49 for the absolute loss function. The algorithm utilizes one random choice at the start of the game. Like the deterministic guess and double algorithm, a deficiency of our algorithm is that it occasionally restarts itself and therefore "forgets" what it learned. Our second algorithm does not forget and enjoys the best known asymptotic performance guarantees for both the absolute and 0/1 loss functions. Specifically, in the case of the absolute loss, our algorithm is canonically bounded with C approaching $$\sqrt 2 $$ and in the case of the 0/1 loss, with C approaching 3/ $$\sqrt 2 \approx 2.12$$ . In the 0/1 loss case the algorithm is randomized and the bound is on the expected regret.