2022/02/23 by Ghoshal, Suprovat, Saha, Aadirupa
#Artificial Intelligence (cs.AI) #FOS: Computer and information sciences #Machine Learning (cs.LG)
paper · doi:10.48550/arxiv.2202.11795
We introduce the Correlated Preference Bandits problem with random utility-based choice models (RUMs), where the goal is to identify the best item from a given pool of n items through online subsetwise preference feedback. We investigate whether models with a simple correlation structure, e.g. low rank, can result in faster learning rates. While we show that the problem can be impossible to solve for the general `low rank' choice models, faster learning rates can be attained assuming more structured item correlations. In particular, we introduce a new class of Block-Rank based RUM model, where the best item is shown to be (ε,δ)-PAC learnable with only O(r ε-2 log(n/δ)) samples. This improves on the standard sample complexity bound of O(nε-2 log(1/δ)) known for the usual learning algorithms which might not exploit the item-correlations (r ≪ n). We complement the above sample complexity with a matching lower bound (up to logarithmic factors), justifying the tightness of our analysis. Surprisingly, we also show a lower bound of Ω(nε-2log(1/δ)) when the learner is forced to play just duels instead of larger subsetwise queries. Further, we extend the results to a more general `noisy Block-Rank' model, which ensures robustness of our techniques. Overall, our results justify the advantage of playing subsetwise queries over pairwise preferences (k=2), we show the latter provably fails to exploit correlation.