ユーザーの知識グラフを用いたLLMによる逐次推薦

抄録

The large language model (LLM) based recommendation system is effective for sequential recommendation and knowledge graph-based prompt tuning has been proposed. However, it is also necessary to consider user knowledge; thus, we propose a user knowledge prompt, which converts a user knowledge graph into a prompt. We conducted experiments on two types of datasets (movie and music) and show the improvement of our proposed user knowledge prompt.