
LLM-ESR is a novel framework that leverages the semantic understanding of Large Language Models (LLMs) to enhance Sequential Recommender Systems (SRS) specifically for long-tail users and items, addressing the challenge of sparse interaction data by incorporating semantic embeddings from LLMs and a novel retrieval augmented self-distillation method.


coremsg

enhancing-long-tailed-sequential-recommendation-with-large-language-models-the-llm-esr-framework


Enhancing Long-Tailed Sequential Recommendation with Large Language Models: The LLM-ESR Framework
