Multi-Item-Query Attention for Stable Sequential Recommendation
This addresses the challenge of unreliable predictions in sequential recommendation for users and platforms, offering an incremental improvement as a drop-in replacement for existing methods.
The paper tackles the problem of instability and noise in sequential recommendation systems by proposing the Multi-Item-Query attention mechanism (MIQ-Attn), which constructs multiple diverse query vectors from user interactions to enhance stability and accuracy, showing significant performance improvements on benchmark datasets.
The inherent instability and noise in user interaction data challenge sequential recommendation systems. Prevailing masked attention models, relying on a single query from the most recent item, are sensitive to this noise, reducing prediction reliability. We propose the Multi-Item-Query attention mechanism (MIQ-Attn) to enhance model stability and accuracy. MIQ-Attn constructs multiple diverse query vectors from user interactions, effectively mitigating noise and improving consistency. It is designed for easy adoption as a drop-in replacement for existing single-query attention. Experiments show MIQ-Attn significantly improves performance on benchmark datasets.