Cross-Domain Meta-Learner for Cold-Start Recommendation

The cold-start problem is a major factor that limits the effectiveness of recommendation systems. Having too few available interaction records brings series challenges when predicting user preferences. At present, there are two main kinds strategies for solving this from different perspectives. One cross-domain (CDR), which introduces additional information by domain knowledge propagation with transfer learning. However, CDR methods follow traditional training processes in machine learning and cannot solve typical few-shot perspective optimization. other type has recently emerged based on meta-learning. Most these approaches focus only generating meta-model to perform better new tasks ignore improvements information. Therefore, it necessary design novel approach both meta-optimization. To achieve goal, meta-learner (MetaCDR) proposed. In MetaCDR, we meta-transfer module connect networks. addition, introduce pretraining strategy ensure its efficiency. experimental results show MetaCDR performs significantly than state-of-the-art models variety scenarios.