Probabilistic Multileave Gradient Descent

Harrie Oosterhuis and Anne Schuth and Maarten de Rijke. In Proceedings of ECIR'16, 2016.

Abstract

Online learning to rank methods aim to optimize ranking models based on user interactions. The dueling bandit gradient descent (DBGD) algorithm is able to effectively optimize linear ranking models solely from user interactions. We propose an extension of DBGD, called probabilistic multileave gradient descent (PMGD) that builds on probabilistic multileave, a recently proposed highly sensitive and unbiased online evaluation method. We demonstrate that P-MGD significantly outperforms state-of-the-art online learning to rank methods in terms of online performance, without sacrificing offline performance and at greater learning speed.

Links

Probabilistic Multileave Gradient Descent
https://doi.org/10.1007/978-3-319-30671-1_50

Bib

@inproceedings{oosterhuis2016,
  title = {Probabilistic Multileave Gradient Descent},
  author = {Harrie Oosterhuis and Anne Schuth and Maarten de Rijke},
  year = {2016},
  booktitle = {Proceedings of ECIR'16},
  doi = {10.1007/978-3-319-30671-1_50}
}