TopRank: A practical algorithm for online stochastic ranking
This work addresses the sequential decision-making challenge in ranking for applications like search engines, offering a more efficient and generalizable solution, though it builds incrementally on prior click models.
The authors tackled the problem of online learning to rank by proposing TopRank, a novel algorithm based on topological sort that operates under a generalized click model, achieving stronger regret guarantees and outperforming existing algorithms empirically.
Online learning to rank is a sequential decision-making problem where in each round the learning agent chooses a list of items and receives feedback in the form of clicks from the user. Many sample-efficient algorithms have been proposed for this problem that assume a specific click model connecting rankings and user behavior. We propose a generalized click model that encompasses many existing models, including the position-based and cascade models. Our generalization motivates a novel online learning algorithm based on topological sort, which we call TopRank. TopRank is (a) more natural than existing algorithms, (b) has stronger regret guarantees than existing algorithms with comparable generality, (c) has a more insightful proof that leaves the door open to many generalizations, (d) outperforms existing algorithms empirically.