Rejuvenating Cross-Entropy Loss in Knowledge Distillation for Recommender Systems
It addresses a specific bottleneck in knowledge distillation for recommender systems, making it incremental but targeted.
This paper tackles the problem of using Cross-Entropy loss in knowledge distillation for recommender systems, where it proves that minimizing CE loss maximizes a lower bound of NDCG only under a closure assumption that conflicts with practical goals, and proposes RCE-KD to bridge this gap, achieving effective results in experiments.
This paper analyzes Cross-Entropy (CE) loss in knowledge distillation (KD) for recommender systems. KD for recommender systems targets at distilling rankings, especially among items most likely to be preferred, and can only be computed on a small subset of items. Considering these features, we reveal the connection between CE loss and NDCG in the field of KD. We prove that when performing KD on an item subset, minimizing CE loss maximizes the lower bound of NDCG, only if an assumption of closure is satisfied. It requires that the item subset consists of the student's top items. However, this contradicts our goal of distilling rankings of the teacher's top items. We empirically demonstrate the vast gap between these two kinds of top items. To bridge the gap between our goal and theoretical support, we propose Rejuvenated Cross-Entropy for Knowledge Distillation (RCE-KD). It splits the top items given by the teacher into two subsets based on whether they are highly ranked by the student. For the subset that defies the condition, a sampling strategy is devised to use teacher-student collaboration to approximate our assumption of closure. We also combine the losses on the two subsets adaptively. Extensive experiments demonstrate the effectiveness of our method. Our code is available at https://anonymous.4open.science/r/RCE-KD.