Retrieval-Oriented Knowledge for Click-Through Rate Prediction

TL;DR

The paper addresses the high inference cost and resource demands of sample-level retrieval-based CTR methods and introduces Retrieval-Oriented Knowledge (ROK), a universal plug-and-play framework that builds a neural knowledge base to imitate retrieved representations via a decomposition-reconstruction paradigm. ROK is optimized with knowledge distillation and contrastive learning to integrate retrieval-enhanced representations with various CTR backbones while enabling retrieval-free inference, achieving $\mathcal{O}(1)$ per-sample complexity. Experiments on three large-scale datasets demonstrate broad compatibility and performance gains across backbones, including surpassing the teacher model, which underscores ROK's potential for real-world deployment by distilling non-parametric retrieval knowledge into a parametric surrogate. The work highlights a practical pathway to transform retrieval-based methods into efficient, scalable industrial solutions for CTR prediction.

Abstract

Click-through rate (CTR) prediction is crucial for personalized online services. Sample-level retrieval-based models, such as RIM, have demonstrated remarkable performance. However, they face challenges including inference inefficiency and high resource consumption due to the retrieval process, which hinder their practical application in industrial settings. To address this, we propose a universal plug-and-play \underline{r}etrieval-\underline{o}riented \underline{k}nowledge (\textbf{\name}) framework that bypasses the real retrieval process. The framework features a knowledge base that preserves and imitates the retrieved \& aggregated representations using a decomposition-reconstruction paradigm. Knowledge distillation and contrastive learning optimize the knowledge base, enabling the integration of retrieval-enhanced representations with various CTR models. Experiments on three large-scale datasets demonstrate \name's exceptional compatibility and performance, with the neural knowledge base serving as an effective surrogate for the retrieval pool. \name surpasses the teacher model while maintaining superior inference efficiency and demonstrates the feasibility of distilling knowledge from non-parametric methods using a parametric approach. These results highlight \name's strong potential for real-world applications and its ability to transform retrieval-based methods into practical solutions. Our implementation code is available to support reproducibility in \url{https://github.com/HSLiu-Initial/ROK.git}.