Inproceedings,

(Bandit) Convex Optimization with Biased Noisy Gradient Oracles

, , , and .
AISTATS, page 819--828. (2016)

Abstract

Algorithms for bandit convex optimization and online learning often rely on constructing noisy gradient estimates, which are then used in appropriately adjusted first-order algorithms, replacing actual gradients. Depending on the properties of the function to be optimized and the nature of ``noise'' in the bandit feedback, the bias and variance of gradient estimates exhibit various tradeoffs. In this paper we propose a novel framework that replaces the specific gradient estimation methods with an abstract oracle. With the help of the new framework we unify previous works, reproducing their results in a clean and concise fashion, while, perhaps more importantly, the framework also allows us to formally show that to achieve the optimal root-n rate either the algorithms that use existing gradient estimators, or the proof techniques used to analyze them have to go beyond what exists today.

Tags

Users

  • @csaba

Comments and Reviews