用极少量参数让固定知识图谱模型个性化，不损失整体性能。
Lightweight Inference-Time Personalization for Frozen Knowledge Graph Embeddings
- 通过用户特征与图结构生成二值门控，动态调整实体偏置。
- 仅需约300个可训练参数，在亚马逊图书和Last-FM数据集上显著提升匹配度。
- 实验证明个性化调整具因果可验证性，适合需要高效个性化的推荐系统。
知识图谱基础模型在链接预测中表现出色，但无法捕捉个体用户偏好，导致通用关系推理与个性化排序之间存在断层。我们提出GatedBias，一种轻量级的推理阶段个性化框架，可在不重训练且不损害全局准确性的前提下，将冻结的知识图谱嵌入适配到具体用户上下文。该方法引入结构门控适应机制：由用户画像特征与图结构衍生的二值门控共同作用，生成可解释的、针对每个实体的偏置，仅需约300个可训练参数。我们在Amazon-Book和Last-FM两个基准数据集上评估了GatedBias，结果显示在对齐指标上具有统计显著提升，同时保持群体性能。反事实扰动实验验证了因果响应性：当特定偏好信号增强时，受益实体的排名提升可达6–30倍。结果表明，基础模型的个性化适配可兼具参数效率与因果可验证性，弥合通用知识表征与个体需求之间的鸿沟。
原文摘要 · Abstract (English)
Foundation models for knowledge graphs (KGs) achieve strong cohort-level performance in link prediction, yet fail to capture individual user preferences; a key disconnect between general relational reasoning and personalized ranking. We propose GatedBias, a lightweight inference-time personalization framework that adapts frozen KG embeddings to individual user contexts without retraining or compromising global accuracy. Our approach introduces structure-gated adaptation: profile-specific features combine with graph-derived binary gates to produce interpretable, per-entity biases, requiring only ${\sim}300$ trainable parameters. We evaluate GatedBias on two benchmark datasets (Amazon-Book and Last-FM), demonstrating statistically significant improvements in alignment metrics while preserving cohort performance. Counterfactual perturbation experiments validate causal responsiveness; entities benefiting from specific preference signals show 6--30$\times$ greater rank improvements when those signals are boosted. These results show that personalized adaptation of foundation models can be both parameter-efficient and causally verifiable, bridging general knowledge representations with individual user needs.
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。