ktexttt-experts -- Online Policies and Fundamental Limits

We introduce the textttk-experts problem - a generalization of the classic Prediction with Expert's Advice framework. Unlike the classic version, where the learner selects exactly one expert from a pool of N experts at each round, in this problem, the learner can select a subset of k experts at each round (1≤ k≤ N). The reward obtained by the learner at each round is assumed to be a function of the k selected experts. The primary objective is to design an online learning policy with a small regret. In this pursuit, we propose textttSAGE (Sampled Hedge) - a framework for designing efficient online learning policies by leveraging statistical sampling techniques. For a wide class of reward functions, we show that textttSAGE either achieves the first sublinear regret guarantee or improves upon the existing ones. Furthermore, going beyond the notion of regret, we fully characterize the mistake bounds achievable by online learning policies for stable loss functions. We conclude the paper by establishing a tight regret lower bound for a variant of the textttk-experts problem and carrying out experiments with standard datasets.