用大模型标注部分候选项，通过高斯过程建模多模态相关性评分，提升自然语言推荐效果。
Multimodal Item Scoring for Natural Language Recommendation via Gaussian Process Regression with LLM Relevance Judgments
- 基于大模型对部分候选项打分，用高斯过程建模多峰相关性分布
- 在4个数据集上相比基线方法最高提升65%，优于点积和余弦相似度
- 仅需少量大模型标注，适合资源有限但追求高精度推荐的场景
自然语言推荐（NLRec）根据用户自然语言请求与物品描述之间的相关性生成推荐。现有方法多使用密集检索（DR）通过用户请求嵌入与相关段落嵌入的内积聚合计算相关性分数，但将请求视为唯一相关性标签，导致评分函数以查询嵌入为中心，难以捕捉复杂语义下的多模态相关性。为此，本文提出GPR-LLM，利用大模型对部分候选段落进行相关性判断，并结合高斯过程回归（GPR）建模潜在的多模态评分分布。在四个NLRec数据集和两种大模型主干上的实验表明，采用RBF核的GPR-LLM能有效建模多模态相关性，显著优于点积、余弦相似度等简单核函数以及基线方法（如DR、cross-encoder、点式大模型评分），最高提升达65%。整体上，GPR-LLM在极小的大模型标注预算下实现高效且精准的自然语言推荐。
原文摘要 · Abstract (English)
Natural Language Recommendation (NLRec) generates item suggestions based on the relevance between user-issued NL requests and NL item description passages. Existing NLRec approaches often use Dense Retrieval (DR) to compute item relevance scores from aggregation of inner products between user request embeddings and relevant passage embeddings. However, DR views the request as the sole relevance label, thus leading to a unimodal scoring function centered on the query embedding that is often a weak proxy for query relevance. To better capture the potential multimodal distribution of the relevance scoring function that may arise from complex NLRec data, we propose GPR-LLM that uses Gaussian Process Regression (GPR) with LLM relevance judgments for a subset of candidate passages. Experiments on four NLRec datasets and two LLM backbones demonstrate that GPR-LLM with an RBF kernel, capable of modeling multimodal relevance scoring functions, consistently outperforms simpler unimodal kernels (dot product, cosine similarity), as well as baseline methods including DR, cross-encoder, and pointwise LLM-based relevance scoring by up to 65%. Overall, GPR-LLM provides an efficient and effective approach to NLRec within a minimal LLM labeling budget.
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。