A Closer Look at Adaptive Regret
For the prediction with expert advice setting, we consider methods to construct algorithms that have low adaptive regret. The adaptive regret of an algorithm on a time interval [t 1,t 2] is the loss of the algorithm there minus the loss of the best expert. Adaptive regret measures how well the algorithm approximates the best expert locally, and it is therefore somewhere between the classical regret (measured on all outcomes) and the tracking regret, where the algorithm is compared to a good sequence of experts.
We investigate two existing intuitive methods to derive algorithms with low adaptive regret, one based on specialist experts and the other based on restarts. Quite surprisingly, we show that both methods lead to the same algorithm, namely Fixed Share, which is known for its tracking regret. Our main result is a thorough analysis of the adaptive regret of Fixed Share. We obtain the exact worst-case adaptive regret for Fixed Share, from which the classical tracking bounds can be derived. We also prove that Fixed Share is optimal, in the sense that no algorithm can have a better adaptive regret bound.
KeywordsOnline learning adaptive regret Fixed Share specialist experts
Unable to display preview. Download preview PDF.
- 2.Cesa-Bianchi, N., Gaillard, P., Lugosi, G., Stoltz, G.: A new look at shifting regret. CoRR abs/1202.3323 (2012)Google Scholar
- 3.Cesa-Bianchi, N., Lugosi, G.: Prediction, Learning, and Games. Cambridge University Press (2006)Google Scholar
- 5.Freund, Y., Schapire, R.E., Singer, Y., Warmuth, M.K.: Using and combining predictors that specialize. In: Proc. 29th Annual ACM Symposium on Theory of Computing, pp. 334–343. ACM (1997)Google Scholar
- 7.Hazan, E., Seshadhri, C.: Efficient learning algorithms for changing environments. In: ICML, p. 50 (2009)Google Scholar
- 9.Koolen, W.M.: Combining Strategies Efficiently: High-quality Decisions from Conflicting Advice. Ph.D. thesis, Institute of Logic, Language and Computation (ILLC), University of Amsterdam (January 2011)Google Scholar
- 10.Koolen, W.M., de Rooij, S.: Combining expert advice efficiently. In: Servedio, R., Zang, T. (eds.) Proceedings of the 21st Annual Conference on Learning Theory (COLT 2008), pp. 275–286 (June 2008)Google Scholar
- 13.Vovk, V.: Aggregating strategies. In: Proceedings of the Third Annual Workshop on Computational Learning Theory, pp. 371–383. Morgan Kaufmann (1990)Google Scholar
- 16.Zinkevich, M.: Online convex programming and generalized infinitesimal gradient ascent. In: Proc. 20th Int. Conference on Machine Learning (ICML 2003), pp. 928–936 (2003)Google Scholar