Timezone: »

 
Poster
Coin Betting and Parameter-Free Online Learning
Francesco Orabona · David Pal

Mon Dec 05 09:00 AM -- 12:30 PM (PST) @ Area 5+6+7+8 #116 #None

In the recent years, a number of parameter-free algorithms have been developed for online linear optimization over Hilbert spaces and for learning with expert advice. These algorithms achieve optimal regret bounds that depend on the unknown competitors, without having to tune the learning rates with oracle choices. We present a new intuitive framework to design parameter-free algorithms for both online linear optimization over Hilbert spaces and for learning with expert advice, based on reductions to betting on outcomes of adversarial coins. We instantiate it using a betting algorithm based on the Krichevsky-Trofimov estimator. The resulting algorithms are simple, with no parameters to be tuned, and they improve or match previous results in terms of regret guarantee and per-round complexity.

Author Information

Francesco Orabona (Yahoo Research)
David Pal (Google)

More from the Same Authors