Oral
Online learning with kernel losses
Niladri S Chatterji · Aldo Pacchiano · Peter Bartlett

Tue Jun 11th 11:40 AM -- 12:00 PM @ Room 103

We present a generalization of the adversarial linear bandits framework, where the underlying losses are kernel functions (with an associated reproducing kernel Hilbert space) rather than linear functions. We study a version of the exponential weights algorithm and bound its regret in this setting. Under conditions on the eigen-decay of the kernel we provide a sharp characterization of the regret for this algorithm. When we have polynomial eigen-decay ($\muj \le \mathcal{O}(j^{-\beta})$), we find that the regret is bounded by $\mathcal{R}n \le \mathcal{O}(n^{\beta/(2\beta-1)})$. While under the assumption of exponential eigen-decay ($\muj \le \mathcal{O}(e^{-\beta j })$) we get an even tighter bound on the regret $\mathcal{R}n \le \tilde{\mathcal{O}}(n^{1/2})$. When the eigen-decay is polynomial we show a \emph{non-matching} minimax lower bound on the regret of $\mathcal{R}n \ge \Omega(n^{(\beta+1)/2\beta})$ and a lower bound of $\mathcal{R}n \ge \Omega(n^{1/2})$ when the decay in the eigen-values is exponentially fast.

We also study the full information setting when the underlying losses are kernel functions and present an adapted exponential weights algorithm and a conditional gradient descent algorithm.