PairRank: Online Pairwise Learning to Rank by Divide-and-Conquer
arXiv:2103.00368 · doi:10.1145/3442381.3449972
Abstract
Online Learning to Rank (OL2R) eliminates the need of explicit relevance annotation by directly optimizing the rankers from their interactions with users. However, the required exploration drives it away from successful practices in offline learning to rank, which limits OL2R's empirical performance and practical applicability. In this work, we propose to estimate a pairwise learning to rank model online. In each round, candidate documents are partitioned and ranked according to the model's confidence on the estimated pairwise rank order, and exploration is only performed on the uncertain pairs of documents, i.e., \emph{divide-and-conquer}. Regret directly defined on the number of mis-ordered pairs is proven, which connects the online solution's theoretical convergence with its expected ranking performance. Comparisons against an extensive list of OL2R baselines on two public learning to rank benchmark datasets demonstrate the effectiveness of the proposed solution.
arXiv commentThe Web Conference 2021
References in corpus (12)
- Introducing LETOR 4.0 Datasets
- Tight Regret Bounds for Stochastic Combinatorial Semi-Bandits
- Cascading Bandits: Learning to Rank in the Cascade Model
- Differentiable Unbiased Online Learning to Rank
- Combinatorial Cascading Bandits
- Neural Contextual Bandits with UCB-based Exploration
- DCM Bandits: Learning to Rank with Multiple Clicks
- Online Learning to Rank in Stochastic Click Models
- TopRank: A practical algorithm for online stochastic ranking
- Variance Reduction in Gradient Exploration for Online Learning to Rank
- Efficient Exploration of Gradient Space for Online Learning to Rank
- Balancing Speed and Quality in Online Learning to Rank for Information Retrieval