79
Voted
NIPS
2007
14 years 11 months ago
2007
We study the rates of growth of the regret in online convex optimization. First, we show that a simple extension of the algorithm of Hazan et al eliminates the need for a priori k...