Leveraging knowledge graphs and contrastive learning to effectively capture user intents across different behaviors and address data sparsity issues for improved multi-behavior recommendation.
