Fusion and Alignment Enhancement with Large Language Models for Tail-item Sequential Recommendation

Researchers propose FAERec, a new framework that uses large language models to improve sequential recommendation systems for rarely-interacted (tail) items. The system addresses fusion and alignment challenges between collaborative signals and semantic knowledge to enhance recommendation accuracy.

• →FAERec framework tackles the tail-item problem in sequential recommendation systems using LLM-derived embeddings.
• →The system introduces an adaptive gating mechanism to dynamically fuse ID and LLM embeddings.
• →Dual-level alignment approach addresses structural inconsistency between ID and LLM embedding spaces.
• →Curriculum learning scheduler prevents premature optimization of complex feature-level objectives.
• →Experiments across three datasets demonstrate effectiveness and generalizability with multiple SR backbones.