从推荐系统中提取可解释的单一语义神经元，实现可控个性化。
Extracting Interaction-Aware Monosemantic Concepts in Recommender Systems
- 用稀疏自编码器发现用户与物品嵌入中的语义结构。
- 通过预测感知目标，保持用户-物品交互一致性，提升语义清晰度。
- 支持内容过滤与推广等操作，适用于多种模型和数据集。
我们提出一种从推荐系统中用户与物品嵌入里提取单义神经元的方法，这些神经元对应于连贯且可解释的概念。该方法采用稀疏自编码器（SAE）揭示预训练表示中的语义结构。不同于语言模型研究，推荐系统中的单义性需保留用户与物品嵌入间的交互关系。为此，我们引入预测感知训练目标，通过冻结的推荐模型反向传播，使学习到的潜在结构与模型的用户-物品亲和度预测对齐。结果神经元捕捉了类型、流行度及时间趋势等属性，并支持事后控制操作，如定向过滤与内容推广，无需修改基础模型。该方法在不同推荐模型与数据集上具有泛化能力，为可解释且可控的个性化提供实用工具。代码与评估资源见 https://github.com/DeltaLabTLV/Monosemanticity4Rec。
原文摘要 · Abstract (English)
We present a method for extracting \emph{monosemantic} neurons, defined as latent dimensions that align with coherent and interpretable concepts, from user and item embeddings in recommender systems. Our approach employs a Sparse Autoencoder (SAE) to reveal semantic structure within pretrained representations. In contrast to work on language models, monosemanticity in recommendation must preserve the interactions between separate user and item embeddings. To achieve this, we introduce a \emph{prediction aware} training objective that backpropagates through a frozen recommender and aligns the learned latent structure with the model's user-item affinity predictions. The resulting neurons capture properties such as genre, popularity, and temporal trends, and support post hoc control operations including targeted filtering and content promotion without modifying the base model. Our method generalizes across different recommendation models and datasets, providing a practical tool for interpretable and controllable personalization. Code and evaluation resources are available at https://github.com/DeltaLabTLV/Monosemanticity4Rec.
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。