Direct Learning to Rank And Rerank

Cynthia Rudin, Yining Wang
Proceedings of the Twenty-First International Conference on Artificial Intelligence and Statistics, PMLR 84:775-783, 2018.

Abstract

Learning-to-rank techniques have proven to be extremely useful for prioritization problems, where we rank items in order of their estimated probabilities, and dedicate our limited resources to the top-ranked items. This work exposes a serious problem with the state of learning-to-rank algorithms, which is that they are based on convex proxies that lead to poor approximations. We then discuss the possibility of "exact" reranking algorithms based on mathematical programming. We prove that a relaxed version of the "exact" problem has the same optimal solution, and provide an empirical analysis.

Cite this Paper


BibTeX
@InProceedings{pmlr-v84-rudin18a, title = {Direct Learning to Rank And Rerank}, author = {Rudin, Cynthia and Wang, Yining}, booktitle = {Proceedings of the Twenty-First International Conference on Artificial Intelligence and Statistics}, pages = {775--783}, year = {2018}, editor = {Storkey, Amos and Perez-Cruz, Fernando}, volume = {84}, series = {Proceedings of Machine Learning Research}, month = {09--11 Apr}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v84/rudin18a/rudin18a.pdf}, url = {https://proceedings.mlr.press/v84/rudin18a.html}, abstract = {Learning-to-rank techniques have proven to be extremely useful for prioritization problems, where we rank items in order of their estimated probabilities, and dedicate our limited resources to the top-ranked items. This work exposes a serious problem with the state of learning-to-rank algorithms, which is that they are based on convex proxies that lead to poor approximations. We then discuss the possibility of "exact" reranking algorithms based on mathematical programming. We prove that a relaxed version of the "exact" problem has the same optimal solution, and provide an empirical analysis. } }
Endnote
%0 Conference Paper %T Direct Learning to Rank And Rerank %A Cynthia Rudin %A Yining Wang %B Proceedings of the Twenty-First International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2018 %E Amos Storkey %E Fernando Perez-Cruz %F pmlr-v84-rudin18a %I PMLR %P 775--783 %U https://proceedings.mlr.press/v84/rudin18a.html %V 84 %X Learning-to-rank techniques have proven to be extremely useful for prioritization problems, where we rank items in order of their estimated probabilities, and dedicate our limited resources to the top-ranked items. This work exposes a serious problem with the state of learning-to-rank algorithms, which is that they are based on convex proxies that lead to poor approximations. We then discuss the possibility of "exact" reranking algorithms based on mathematical programming. We prove that a relaxed version of the "exact" problem has the same optimal solution, and provide an empirical analysis.
APA
Rudin, C. & Wang, Y.. (2018). Direct Learning to Rank And Rerank. Proceedings of the Twenty-First International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 84:775-783 Available from https://proceedings.mlr.press/v84/rudin18a.html.

Related Material