CS3: Efficient Online Capability Synergy for Two-Tower Recommendation

• URL: http://arxiv.org/abs/2604.19269v1
• Date: Tue, 21 Apr 2026 09:35:30 GMT
• Title: CS3: Efficient Online Capability Synergy for Two-Tower Recommendation
• Abstract summary: Two-tower architecture limits representation capacity, embedding-space alignment, and cross-feature interactions.<n>We propose Capability Synergy (CS3), an efficient online framework that strengthens two-tower retrievers while preserving real-time constraints.<n> CS3 is plug-and-play with diverse two-tower backbones and compatible with online learning.
• Score: 8.955059715857173
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: To balance effectiveness and efficiency in recommender systems, multi-stage pipelines commonly use lightweight two-tower models for large-scale candidate retrieval. However, the isolated two-tower architecture restricts representation capacity, embedding-space alignment, and cross-feature interactions. Existing solutions such as late interaction and knowledge distillation can mitigate these issues, but often increase latency or are difficult to deploy in online learning settings. We propose Capability Synergy (CS3), an efficient online framework that strengthens two-tower retrievers while preserving real-time constraints. CS3 introduces three mechanisms: (1) Cycle-Adaptive Structure for self-revision via adaptive feature denoising within each tower; (2) Cross-Tower Synchronization to improve alignment through lightweight mutual awareness between towers; and (3) Cascade-Model Sharing to enhance cross-stage consistency by reusing knowledge from downstream models. CS3 is plug-and-play with diverse two-tower backbones and compatible with online learning. Experiments on three public datasets show consistent gains over strong baselines, and deployment in a largescale advertising system yields up to 8.36% revenue improvement across three scenarios while maintaining ms-level latency.

Related papers

• MambaADv2: Evolving Duality-enhanced State Space Model for Unsupervised Anomaly Detection [108.36437360254605]
  This paper proposes MambaADv2, a framework tailored for multi-class unsupervised anomaly detection.<n>MambaADv2 comprises a pre-trained encoder and a Mamba-inspired decoder, equipped with Duality-enhanced State Space (DSS) modules across multiple scales.<n>The structure of the Hybrid State Space (HSS) block is tailored by following the SSD-based Mamba lineage and incorporating Mamba3-style position-aware state-space modeling.
  arXiv  Detail & Related papers  (2026-06-22T10:14:06Z)
• CS3: Efficient Online Capability Synergy for Two-Tower Recommendation [8.955059715857173]
  We propose an efficient online framework called Capability Synergy (CS3) to enhance two-tower models.<n>The CS3 framework is compatible with various two-tower architectures and meets real-time requirements in online learning scenarios.<n> Experimental results demonstrate that CS3 increases online ad revenue by up to 8.36% across three scenarios.
  arXiv  Detail & Related papers  (2026-03-10T08:25:53Z)
• Efficient Personalized Reranking with Semi-Autoregressive Generation and Online Knowledge Distillation [67.09494009022471]
  We propose a novel Personalized Semi-Autoregressive with online knowledge Distillation (PSAD) framework for reranking.<n>In this framework, the teacher model adopts a semi-autoregressive generator to balance generation quality and efficiency, while its ranking knowledge is distilled online into a lightweight scoring network.<n>Experiments conducted on three large-scale public datasets demonstrate that PSAD significantly outperforms state-of-the-art baselines in both ranking performance and inference efficiency.
  arXiv  Detail & Related papers  (2026-03-07T08:37:13Z)
• CATNet: Collaborative Alignment and Transformation Network for Cooperative Perception [9.983779569276475]
  Collaborative Alignment and Transformation Network (CATNet) is an adaptive compensation framework that resolves temporal latency and noise interference in multi-agent systems.<n>Our key innovations can be summarized in three aspects. First, we introduce a Spatio-Temporal Recurrent Synchronization (STSync) that aligns asynchronous feature streams.<n>Second, we design a Dual-Branch Wavelet Enhanced Denoiser (WTDen) that suppresses global noise and reconstructs localized feature distortions.<n>Third, we construct an Adaptive Feature Selector (AdpSel) that dynamically focuses on critical perceptual features for robust fusion.
  arXiv  Detail & Related papers  (2026-03-05T15:07:36Z)
• Compress, Cross and Scale: Multi-Level Compression Cross Networks for Efficient Scaling in Recommender Systems [5.897678894426804]
  MLCC is a structured feature interaction architecture that organizes feature crosses through hierarchical compression and dynamic composition.<n>MC-MLCC is a Multi-Channel extension that decomposes feature interactions into parallel subspaces.<n>Our proposed models consistently outperform strong DLRM-style baselines by up to 0.52 AUC, while reducing model parameters and FLOPs by up to 26$times$ under comparable performance.
  arXiv  Detail & Related papers  (2026-02-12T15:06:46Z)
• A2R: An Asymmetric Two-Stage Reasoning Framework for Parallel Reasoning [57.727084580884075]
  Asymmetric Two-Stage Reasoning framework designed to bridge gap between a model's potential and its actual performance.<n>A2R-Efficient is a "small-to-big" variant that combines a Qwen3-4B explorer with a Qwen3-8B synthesizer.<n>Results show A2R is not only a performance-boosting framework but also an efficient and practical solution for real-world applications.
  arXiv  Detail & Related papers  (2025-09-26T08:27:03Z)
• C2D-ISR: Optimizing Attention-based Image Super-resolution from Continuous to Discrete Scales [6.700548615812325]
  We propose a novel framework, textbfC2D-ISR, for optimizing attention-based image super-resolution models.<n>Our approach is based on a two-stage training methodology and a hierarchical encoding mechanism.<n>In addition, we generalize the hierarchical encoding mechanism with existing attention-based network structures.
  arXiv  Detail & Related papers  (2025-03-17T21:52:18Z)
• Dynamic Cross-Modal Feature Interaction Network for Hyperspectral and LiDAR Data Classification [66.59320112015556]
  Hyperspectral image (HSI) and LiDAR data joint classification is a challenging task.<n>We propose a novel Dynamic Cross-Modal Feature Interaction Network (DCMNet)<n>Our approach introduces three feature interaction blocks: Bilinear Spatial Attention Block (BSAB), Bilinear Channel Attention Block (BCAB), and Integration Convolutional Block (ICB)
  arXiv  Detail & Related papers  (2025-03-10T05:50:13Z)
• Climber: Toward Efficient Scaling Laws for Large Recommendation Models [8.970144942471425]
  We propose Climber, an efficient recommendation framework comprising two synergistic components.<n>Our proposed model adopts two core innovations: (1) multi-scale sequence extraction that achieves a time complexity reduction by a constant factor, enabling more efficient scaling with sequence length; (2) dynamic temperature modulation adapting attention distributions to the multi-scenario and multi-behavior patterns.<n> Climber has been successfully deployed on Netease Cloud Music, one of China's largest music streaming platforms, serving tens of millions of users daily.
  arXiv  Detail & Related papers  (2025-02-14T03:25:09Z)
• FusionLLM: A Decentralized LLM Training System on Geo-distributed GPUs with Adaptive Compression [55.992528247880685]
  Decentralized training faces significant challenges regarding system design and efficiency. We present FusionLLM, a decentralized training system designed and implemented for training large deep neural networks (DNNs) We show that our system and method can achieve 1.45 - 9.39x speedup compared to baseline methods while ensuring convergence.
  arXiv  Detail & Related papers  (2024-10-16T16:13:19Z)
• Mutual Information-driven Triple Interaction Network for Efficient Image Dehazing [54.168567276280505]
  We propose a novel Mutual Information-driven Triple interaction Network (MITNet) for image dehazing. The first stage, named amplitude-guided haze removal, aims to recover the amplitude spectrum of the hazy images for haze removal. The second stage, named phase-guided structure refined, devotes to learning the transformation and refinement of the phase spectrum.
  arXiv  Detail & Related papers  (2023-08-14T08:23:58Z)
• VFed-SSD: Towards Practical Vertical Federated Advertising [53.08038962443853]
  We propose a semi-supervised split distillation framework VFed-SSD to alleviate the two limitations. Specifically, we develop a self-supervised task MatchedPair Detection (MPD) to exploit the vertically partitioned unlabeled data. Our framework provides an efficient federation-enhanced solution for real-time display advertising with minimal deploying cost and significant performance lift.
  arXiv  Detail & Related papers  (2022-05-31T17:45:30Z)
• Dynamic Scheduling for Stochastic Edge-Cloud Computing Environments using A3C learning and Residual Recurrent Neural Networks [30.61220416710614]
  A-Advantage-Actor-Critic (A3C) learning is known to quickly adapt to dynamic scenarios with less data and Residual Recurrent Neural Network (R2N2) to quickly update model parameters. We use the R2N2 architecture to capture a large number of host and task parameters together with temporal patterns to provide efficient scheduling decisions. Experiments conducted on real-world data set show a significant improvement in terms of energy consumption, response time, ServiceLevelAgreement and running cost by 14.4%, 7.74%, 31.9%, and 4.64%, respectively.
  arXiv  Detail & Related papers  (2020-09-01T13:36:34Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.