用稀疏路由+概率存储，让多模态推荐更省空间且效果更好
SA-RSQ: A Versatile Sparse Representation Framework for Multi-modal Recommender Systems
- 通过顶K稀疏路由和软量化权重，存储(索引, 概率)元组
- 在8-48字节/项下，重建精度与点击率平衡更优
- 适合工业级推荐系统，实测点击率提升2.51%，广告收益升3.66%
工业级推荐系统中部署高维多模态特征会带来显著的存储与延迟开销。硬量化虽紧凑但存在边界失真，而密集软量化则将表示质量受限于存储预算。我们提出基于稀疏激活的残差软量化（SA-RSQ），采用顶K稀疏路由与softmax权重，存储紧凑的(索引, 概率)元组。该方法使每项存储量与码本维度解耦；固定选中支持下，梯度通过路由权重与加权重构传播，无需直通估计器。在自有外卖广告数据集上的实验表明，在8-48字节/项的存储预算下，该方法在重建性能与点击率（CTR）间取得良好权衡。初步的下一分布预测研究及一周在线A/B测试进一步验证其实用性，分别实现+2.51% CTR与+3.66% CPM的相对提升。
原文摘要 · Abstract (English)
Deploying high-dimensional multimodal features in industrial recommender systems incurs substantial storage and latency overhead. Hard quantization is compact but introduces boundary distortion, whereas dense soft quantization couples representation quality to the limited storage budget. We propose Sparse Activation-based Residual Soft Quantization (SA-RSQ), which uses Top-K sparse routing and softmax weights to store compact (Index, Probability) tuples. The stored tuples decouple per-item storage from codebook dimensionality; for a fixed selected support, gradients propagate through the routing weights and weighted reconstruction without relying on a straight-through estimator. Experiments on a proprietary food-delivery advertising dataset show favorable reconstruction-performance and CTR trade-offs across storage budgets of 8-48 bytes per item. A preliminary Next-Distribution Prediction study and a one-week online A/B test further demonstrate the practical potential of SA-RSQ, with relative lifts of +2.51% in CTR and +3.66% in CPM.
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。