Triplet Losses-based Matrix Factorization for Robust Recommendations
This work addresses bias issues in recommender systems that affect certain user groups, offering an incremental improvement over existing methods.
The paper tackled bias in recommender systems by proposing a matrix factorization method using multiple triplet losses to create robust user and item representations, and evaluated it with bias-aware metrics, showing improved stability and prediction agreement.
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.