ML Anthology

Computationally Efficient Optimization of Plackett-Luce Ranking Models for Relevance and Fairness (Extended Abstract)

Harrie Oosterhuis
IJCAI 2022 pp. 5319-5323
doi:10.24963/IJCAI.2022/743 /ijcai/2022/oosterhuis2022ijcai-computationally/

Abstract

Computing the gradient of stochastic Plackett-Luce (PL) ranking models for relevance and fairness metrics can be infeasible because it requires iterating over all possible permutations of items. In this paper, we introduce a novel algorithm: PL-Rank, that estimates the gradient of a PL ranking model through sampling. Unlike existing approaches, PL-Rank makes use of the specific structure of PL models and ranking metrics. Our experimental analysis shows that PL-Rank has a greater sample-efficiency and is computationally less costly than existing policy gradients, resulting in faster convergence at higher performance.

PDF IJCAI Semantic Scholar

Cite

Text

Oosterhuis. "Computationally Efficient Optimization of Plackett-Luce Ranking Models for Relevance and Fairness (Extended Abstract)." International Joint Conference on Artificial Intelligence, 2022. doi:10.24963/IJCAI.2022/743

Markdown

[Oosterhuis. "Computationally Efficient Optimization of Plackett-Luce Ranking Models for Relevance and Fairness (Extended Abstract)." International Joint Conference on Artificial Intelligence, 2022.](https://mlanthology.org/ijcai/2022/oosterhuis2022ijcai-computationally/) doi:10.24963/IJCAI.2022/743

BibTeX

@inproceedings{oosterhuis2022ijcai-computationally,
  title     = {{Computationally Efficient Optimization of Plackett-Luce Ranking Models for Relevance and Fairness (Extended Abstract)}},
  author    = {Oosterhuis, Harrie},
  booktitle = {International Joint Conference on Artificial Intelligence},
  year      = {2022},
  pages     = {5319-5323},
  doi       = {10.24963/IJCAI.2022/743},
  url       = {https://mlanthology.org/ijcai/2022/oosterhuis2022ijcai-computationally/}
}