ListMAP: Listwise learning to rank as maximum a posteriori estimation

作者：

Highlights：

• ListMAP, a new listwise learning to rank model with prior distribution to weight training instances, is introduced.

• A model for approximating the prior distribution parameters from a set of observation data is introduced.

• The characteristics of the introduced L2R models with respect to coherency are discussed and analyzed.

• The proposed model outperforms the prominent and state-of-the-art leaning to rank models across different datasets.

摘要

•ListMAP, a new listwise learning to rank model with prior distribution to weight training instances, is introduced.•A model for approximating the prior distribution parameters from a set of observation data is introduced.•The characteristics of the introduced L2R models with respect to coherency are discussed and analyzed.•The proposed model outperforms the prominent and state-of-the-art leaning to rank models across different datasets.

论文关键词：Learning to rank,Listwise loss function,Prior distribution

论文评审过程：Received 9 November 2021, Revised 28 February 2022, Accepted 24 April 2022, Available online 13 May 2022, Version of Record 13 May 2022.

论文官网地址：https://doi.org/10.1016/j.ipm.2022.102962