English

Towards Non-Parametric Learning to Rank

Machine Learning 2018-07-11 v1 Artificial Intelligence Computers and Society Machine Learning

Abstract

This paper studies a stylized, yet natural, learning-to-rank problem and points out the critical incorrectness of a widely used nearest neighbor algorithm. We consider a model with nn agents (users) {xi}i[n]\{x_i\}_{i \in [n]} and mm alternatives (items) {yj}j[m]\{y_j\}_{j \in [m]}, each of which is associated with a latent feature vector. Agents rank items nondeterministically according to the Plackett-Luce model, where the higher the utility of an item to the agent, the more likely this item will be ranked high by the agent. Our goal is to find neighbors of an arbitrary agent or alternative in the latent space. We first show that the Kendall-tau distance based kNN produces incorrect results in our model. Next, we fix the problem by introducing a new algorithm with features constructed from "global information" of the data matrix. Our approach is in sharp contrast to most existing feature engineering methods. Finally, we design another new algorithm identifying similar alternatives. The construction of alternative features can be done using "local information," highlighting the algorithmic difference between finding similar agents and similar alternatives.

Keywords

Cite

@article{arxiv.1807.03395,
  title  = {Towards Non-Parametric Learning to Rank},
  author = {Ao Liu and Qiong Wu and Zhenming Liu and Lirong Xia},
  journal= {arXiv preprint arXiv:1807.03395},
  year   = {2018}
}

Comments

10 pages' main document and 10 pages' supplementary documents

R2 v1 2026-06-23T02:55:39.389Z