CS3: Efficient Online Capability Synergy for Two-Tower Recommendation

TL;DR

The paper introduces CS3, an online framework that enhances two-tower recommendation models by improving their representation and alignment capabilities without sacrificing real-time performance.

Contribution

CS3 provides a novel, efficient online enhancement for two-tower models through cycle-adaptive structure, cross-tower synchronization, and cascade model sharing.

Findings

CS3 increases online ad revenue by up to 8.36%.

Maintains millisecond-level latency in large-scale deployment.

Compatible with various two-tower architectures.

Abstract

To balance effectiveness and efficiency in recommender systems, multi-stage pipelines employ lightweight two-tower models for large-scale candidate retrieval. However, their isolated architecture inherently hampers representation capacity, embedding-space alignment, and cross-feature modeling. Prior studies have explored incorporating late interaction or knowledge distillation to mitigate these issues, but such approaches often significantly increase model latency or pose challenges for implementation in online learning scenarios. To address these limitations, we propose an efficient online framework called Capability Synergy (CS3), which enhances two-tower models through three key innovations: (1) Cycle-Adaptive Structure, enabling self-revision via adaptive feature denoising within individual towers; (2) Cross-Tower Synchronization, improving representation alignment through mutual awareness between the towers; and (3) CascadeModel Sharing, bridging cross-stage consistency by reusing knowledge from downstream models. The CS3 framework is compatible with various two-tower architectures and meets real-time requirements in online learning scenarios. We evaluated CS3 on three public offline datasets and subsequently deployed it in a large-scale advertising system. Experimental results demonstrate that CS3 increases online ad revenue by up to 8.36% across three scenarios while maintaining millisecond-level latency and consistently performing well across diverse two-tower architectures.