RIA: A Ranking-Infused Approach for Optimized listwise CTR Prediction

TL;DR

The paper tackles the challenge of estimating CTR with strong listwise quality under tight latency by bridging pointwise and listwise evaluation in a single end-to-end framework. It introduces RIA, featuring four modules—UCDT for fine-grained user-item-context interactions, CUHT with PIAU and PTAU for position-aware history learning, LMH for hierarchical list dependencies, and EC for efficient inference—sharing representations to enable robust knowledge transfer. Empirical results show consistent offline improvements in AUC and LogLoss on public (Avito) and industrial (Meituan) data, and online A/B tests report CTR gains up to +2.11% and CPM gains up to +5.83% with modest latency increases. The work demonstrates the practical value of a unified ranking-reranking paradigm in large-scale advertising, offering both performance enhancements and deployment efficiency, and it validates the approach through real-world Meituan deployment.

Abstract

Reranking improves recommendation quality by modeling item interactions. However, existing methods often decouple ranking and reranking, leading to weak listwise evaluation models that suffer from combinatorial sparsity and limited representational power under strict latency constraints. In this paper, we propose RIA (Ranking-Infused Architecture), a unified, end-to-end framework that seamlessly integrates pointwise and listwise evaluation. RIA introduces four key components: (1) the User and Candidate DualTransformer (UCDT) for fine-grained user-item-context modeling; (2) the Context-aware User History and Target (CUHT) module for position-sensitive preference learning; (3) the Listwise Multi-HSTU (LMH) module to capture hierarchical item dependencies; and (4) the Embedding Cache (EC) module to bridge efficiency and effectiveness during inference. By sharing representations across ranking and reranking, RIA enables rich contextual knowledge transfer while maintaining low latency. Extensive experiments show that RIA outperforms state-of-the-art models on both public and industrial datasets, achieving significant gains in AUC and LogLoss. Deployed in Meituan advertising system, RIA yields a +1.69% improvement in Click-Through Rate (CTR) and a +4.54% increase in Cost Per Mille (CPM) in online A/B tests.

Figures (4)

• Figure 1: An illustration of combinatorial sparsity.
• Figure 2: Overview of our proposed method ($m=3$).
• Figure 3: Architecture of CUHT ($m=3$).
• Figure 4: AUC vs. Number of HSTU layers.