Triplet Losses-based Matrix Factorization for Robust Recommendations

Flavio Giobergia

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 t…