QuickScorer: a Fast Algorithm to Rank Documents with Additive Ensembles of Regression Trees
Accepted at the 38th International ACM SIGIR Conference on Research and Development in Information Retrieval, Santiago de Chile, 9–13 August 2015 .
Abstract. Learning-to-Rank models based on additive ensembles of regression trees have proven to be very eﬀective for ranking query results returned by Web search engines, a scenario where quality and eﬃciency requirements are very demanding. Unfortunately, the computational cost of these ranking models is high. Thus, several works already proposed solutions aiming at improving the eﬃciency of the scoring process by dealing with features and peculiarities of modern CPUs and memory hierarchies. In this paper, we present QuickScorer, a new algorithm that adopts a novel bitvector representation of the tree-based ranking model, and performs an interleaved traversal of the ensemble by means of simple logical bitwise operations. The performance of the proposed algorithm are unprecedented, due to its cache-aware approach, both in terms of data layout and access patterns, and to a control ﬂow that entails very low branch mis-prediction rates. The experiments on real Learning-to-Rank datasets show that QuickScorer is able to achieve speedups over the best state-of-the-art baseline ranging from 2x to 6.5x.
 Claudio Lucchese, Franco Maria Nardini, Salvatore Orlando, Raﬀaele Perego, Nicola Tonellotto, and Rossano Venturini. Quickscorer: a fast algorithm to rank documents with additive ensembles of regression trees. In SIGIR ’15: Proceedings of the 38th International ACM SIGIR Conference on Research and Development in Information Retrieval, 2015. (best paper) (ACM Notable Article).