Use this link to cite:

https://hdl.handle.net/2183/48855

Contrastive Learning for Explanation Ranking

Thumbnail Image

Files

Cancela_Brais_2026_Contrastive_Learning_for_Explanation_Ranking.pdf (2.82 MB)

Identifiers

URI: https://hdl.handle.net/2183/48855
DOI: 10.1007/s10994-026-07099-7

Publication date

2026

Authors

Escarda Fernández, Miguel

Advisors

Other responsabilities

Journal Title

Bibliographic citation

Escarda-Fernández, M., Cancela, B., Eiras-Franco, C. et al. Contrastive Learning for Explanation Ranking. Mach Learn 115, 156 (2026). https://doi.org/10.1007/s10994-026-07099-7

Type of academic work

Academic degree

Abstract

[Abstract]: Explainable recommendation systems enhance user trust and satisfaction by revealing the reasoning behind personalized recommendations. Approaching this as a post-hoc explanation-ranking problem over a fixed pool of candidate explanations, we propose Contrastive Learning for Explanation Ranking (CLER), a model that learns user, item, and explanation representations with a Normalized Temperature-scaled Binary Cross-Entropy (NT-BXent) loss. This function specifically applies a per-row reweighting strategy, preventing the vast number of negative examples from dominating the objective. We evaluate CLER on the Amazon, TripAdvisor, and Yelp datasets from the EXTRA benchmark. Across traditional ranking metrics, CLER achieves the strongest results among the compared baselines.

Description

Financiado para publicación en acceso aberto: CRUE-CSIC

Keywords

Learning-to-rank Explanation recommendation Contrastive learning Post-hoc explanations

Editor version

https://doi.org/10.1007/s10994-026-07099-7

Rights

Attribution 4.0 International

Collections

Investigación (FIC)
Attribution 4.0 International

Except where otherwise noted, this item's license is described as Attribution 4.0 International