P. Auer, N. Cesa-Bianchi, and C. Gentile. Adaptive and self-confident on-line learning algorithms. Journal of Computer and System Sciences, 64:48--75, 2002.  Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Smooth Online Learning of Expert Advice - Yaroshinsky, El-Yaniv (2001) (2 citations) (Correct)
....to be bounded above by L # (3.32 o(1) # L # ln N . 2.4 Algorithm iAWM The major deficiency of the doubling algorithm P # is the initialization it performs at the start of each phase where it forgets all the information it learned about the experts. Recently, Auer, Cesa Bianchi and Gentile [ACBG00] proposed a new self confident algorithm, called iAWM (stands for Incrementally Adaptive Weighted Majority) which incrementally updates the learning rate and does not forget expert weights. The (absolute loss) regret of iAWM is bounded above by (2.83 o(1) # L # ln N . The algorithm s ....
...., j = 1, N . L # i 1 min 1#j#N (L j,i 1 ) min L # i 1 . Receive experts predictions # . Predict y = w i . Punish experts weights: # N : L j,i L j,i y y . Figure 4: The Incrementally Adaptive Weighted Majority Algorithm (iAWM) of [ACBG00] 2.5 P norm algorithms Gentile and Littlestone considered P norm algorithms in the context of expert advice. They show that for the 0 1 loss, the mistake bound of P norm algorithms is bounded above by 2 L # (e N . A regret bound of L # (e The non asymptotic ....
[Article contains additional citation context not shown here]
P. Auer, N. Ceza-Bianchi, and C. Gentile. Adaptive and self-confident on-line learning algorithms. NeuroCOLT2 Technical Report Series, November 2000. NC2-TR-2000083 (Also apeared in the Proc. of COLT2000).
Smooth Online Learning of Expert Advice - Yaroshinsky, El-Yaniv (2001) (2 citations) (Correct)
....be bounded above by L # (3.32 o(1) # L # ln N . 4 2.4 Algorithm iAWM The major deficiency of the doubling algorithm P # is the initialization it performs at the start of each phase where it forgets all the information it learned about the experts. Recently, Auer, Cesa Bianchi and Gentile [ACBG00] proposed a new self confident algorithm, called iAWM (stands for Incrementally Adaptive Weighted Majority) which incrementally updates the learning rate and does not forget expert weights. The (absolute loss) regret of iAWM is bounded above by (2.83 o(1) # L # ln N . 5 The algorithm s ....
.... . # i # 1 1 # i . Receive experts predictions # # [0, 1] N . Predict y = w i # . Receive true outcome y # [0, 1] Punish experts weights: # 1 # j # N : L j,i # L j,i y y . endfor Figure 4: The Incrementally Adaptive Weighted Majority Algorithm (iAWM) of [ACBG00] 2.5 P norm algorithms Gentile and Littlestone considered P norm algorithms in the context of expert advice. 6 They show that for the 0 1 loss, the mistake bound of P norm algorithms is bounded above by 2 L # (e 2) ln N 2 # e # L # ln N e 4 ln 2 N . A regret bound of L ....
[Article contains additional citation context not shown here]
P. Auer, N. Ceza-Bianchi, and C. Gentile. Adaptive and self-confident on-line learning algorithms. NeuroCOLT2 Technical Report Series, November
Journal of Machine Learning Research 7 (2006) 1205--1230 .. - For Linear.. Self-citation (Cesa-bianchi Gentile) (Correct)
No context found.
P. Auer, N. Cesa-Bianchi, and C. Gentile. Adaptive and self-confident on-line learning algorithms. Journal of Computer and System Sciences, 64:48--75, 2002.
Optimality of Universal Bayesian Sequence Prediction for General.. - Hutter (2003) (1 citation) (Correct)
No context found.
P. Auer and C. Gentile. Adaptive and self-confident on-line learning algorithms. In Proceedings of the 13th Conference on Computational Learning Theory, pages 107--117. Morgan Kaufmann, San Francisco, 2000.
Unknown - (Correct)
No context found.
P. Auer, N. Cesa-Bianchi, and C. Gentile. Adaptive and self-confident on-line learning algorithms. Journal of Computer and System Sciences, 64(1):48--75, 2002.
Online Learning with Kernels - Kivinen, Smola, Williamson (2002) (5 citations) (Correct)
No context found.
P. Auer, N. Cesa-Bianchi, and C. Gentile, "Adaptive and self-confident on-line learning algorithms," Journal of Computer and System Sciences, vol. 64, no. 1, pp. 48--75, Feb. 2002.
Large Margin Classification for Moving Targets - Kivinen, Smola, Williamson (2001) (Correct)
No context found.
P. Auer, N. Cesa-Bianchi and C. Gentile. Adaptive and self-confident on-line learning algorithms. Technical Report NC-TR-00-083, NeuroCOLT, 2000.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC