不训练模型，用协同信号精炼物品表示，提升推荐精度
Training-Free LLM-Based Recommendation with Post-LLM Item Refinement Using Collaborative Signals
- 在LLM生成物品表示后，注入共购图和流行度信息进行优化
- 在多个真实数据集上超越现有免训练方法，媲美有训练模型
- 适合追求零训练成本、快速部署的推荐系统场景
大语言模型（LLM）在免训练推荐中展现出潜力，但其生成的用户兴趣通常过于宽泛，难以实现细粒度物品检索。现有方法通过候选重排或提示增强在预LLM阶段融合协同过滤（CF）信号，效果有限。本文提出CoRRe，一种基于后LLM范式的免训练推荐框架，将CF信号注入LLM生成的物品表示中，并与用户兴趣匹配排序。具体地，CoRRe利用物品-物品共购图调整物品嵌入方向，用物品流行度调节其幅值。在真实数据集上的实验表明，CoRRe持续优于现有免训练方法，在无需任何模型训练或任务微调的情况下，达到与训练型方法相当甚至更优的性能。
原文摘要 · Abstract (English)
Large language models (LLMs) have shown promise for training-free recommendation, but LLM-generated user interests are often too broad for fine-grained item retrieval. Existing methods incorporate collaborative filtering (CF) signals in a pre-LLM manner through candidate reranking or prompt augmentation, yielding limited gains. We propose CoRRe, a training-free recommendation framework with a post-LLM paradigm that injects CF signals into LLM-generated item representations, which are later matched with LLM-generated user interests for ranking. Specifically, CoRRe refines the directions of item embeddings using an item-item co-purchase graph and their magnitudes using item popularity. Experiments on real-world datasets show that CoRRe consistently outperforms existing training-free methods and achieves competitive or superior performance compared with training-based methods, without requiring any model training or task-specific fine-tuning.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。