Online Learning to Rank in Stochastic Click Models
This addresses the practical limitation of existing online learning to rank algorithms for information retrieval by providing a more general solution, though it is incremental in extending coverage to multiple click models.
The paper tackled the problem of online learning to rank algorithms being limited to specific click models by proposing BatchRank, the first algorithm for a broad class including cascade and position-based models, and showed it outperforms ranked bandits and is more robust than CascadeKL-UCB in evaluations on web search queries.
Online learning to rank is a core problem in information retrieval and machine learning. Many provably efficient algorithms have been recently proposed for this problem in specific click models. The click model is a model of how the user interacts with a list of documents. Though these results are significant, their impact on practice is limited, because all proposed algorithms are designed for specific click models and lack convergence guarantees in other models. In this work, we propose BatchRank, the first online learning to rank algorithm for a broad class of click models. The class encompasses two most fundamental click models, the cascade and position-based models. We derive a gap-dependent upper bound on the $T$-step regret of BatchRank and evaluate it on a range of web search queries. We observe that BatchRank outperforms ranked bandits and is more robust than CascadeKL-UCB, an existing algorithm for the cascade model.