Oracle-Efficient Hybrid Online Learning with Unknown Distribution

We study the problem of oracle-efficient hybrid online learning when the features are generated by an unknown i.i.d. process and the labels are generated adversarially. Assuming access to an (offline) ERM oracle, we show that there exists a computationally efficient online predictor that achieves a regret upper bounded by Õ(T^3/4) for a finite-VC class, and upper bounded by Õ(T^p+1/p+2) for a class with α fat-shattering dimension α^-p. This provides the first known oracle-efficient sublinear regret bounds for hybrid online learning with an unknown feature generation process. In particular, it confirms a conjecture of Lazaric and Munos (JCSS 2012). We then extend our result to the scenario of shifting distributions with K changes, yielding a regret of order Õ(T^4/5K^1/5). Finally, we establish a regret of Õ((K^2/3(log|ℋ|)^1/3+K)· T^4/5) for the contextual K-armed bandits with a finite policy set ℋ, i.i.d. generated contexts from an unknown distribution, and adversarially generated costs.