Triplet Losses-based Matrix Factorization for Robust Recommendations

TL;DR

This paper introduces a novel matrix factorization approach using triplet losses to create more robust and bias-aware user and item representations in recommender systems, improving stability and fairness.

Contribution

It proposes a new triplet loss-based matrix factorization method that enhances robustness and bias-awareness in recommendations, validated through comprehensive bias-aware evaluations.

Findings

Improved stability of recommendations against training data changes

Enhanced bias mitigation in user-item representations

Better alignment of prediction variance with user-specific behaviors

Abstract

Much like other learning-based models, recommender systems can be affected by biases in the training data. While typical evaluation metrics (e.g. hit rate) are not concerned with them, some categories of final users are heavily affected by these biases. In this work, we propose using multiple triplet losses terms to extract meaningful and robust representations of users and items. We empirically evaluate the soundness of such representations through several "bias-aware" evaluation metrics, as well as in terms of stability to changes in the training set and agreement of the predictions variance w.r.t. that of each user.