mlsanityData to 7 Oct 2026

Heterogeneous Ranking in Industrial-Scale Recommender Systems: A Case Study

arXiv 2607.27577 ↗First posted 30 Jul 2026cs.IR cs.LGcited by 0

On the map: Mixture-of-experts → MoE for recommender systems

Abstract (arXiv)

Heterogeneous recommendation feeds present complex challenges that extend beyond those found in highly homogeneous environments (e.g., music-only or video-only closed-ecosystem platforms). In Google Discover, a unified feed integrates diverse content sourced from the decentralized open web, including web articles, long-form and short-form videos, user-generated content (UGC), and beyond. Different content types exhibit distinct feature densities and user interaction patterns. Building a unified ranking model that sustains high performance across such heterogeneity, while avoiding negative transfer or majority bias, remains a significant industrial challenge. This paper presents an end-to-end case study on the industrial-scale multi-task ranking of heterogeneous feeds, grounded in real-world deployment. We introduce HA-MoE, a heterogeneity-adaptive multi-gated mixture-of-experts architecture that incorporates explicit heterogeneity context into both gating networks and expert representations. This approach enables effective specialization without significantly increasing operational overhead. To support reliable deployment, we introduce LENS, a lightweight observability framework that provides interpretable diagnostics of expert specialization and tracks this functional heterogeneity across continuous retraining. We evaluate our method using Dual-Level AUC (DL-AUC), a heterogeneity-aware evaluation metric that combines global ranking performance with cross-segment ranking correctness. Offline evaluations on a large-scale industrial dataset demonstrate consistent improvements over baseline models. Furthermore, online A/B testing confirms gains in feed activity and exploration metrics. Together, offline and online results validate the effectiveness of our approach for managing heterogeneity in industrial-scale recommender systems.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in MoE for recommender systems, one of 12 directions in Mixture-of-experts. Builds on 22 of its 38 references that are in the map. First posted on arXiv 30 Jul 2026.

Builds on

22 of 38 references.

Closest ancestors · 1

Same direction, newest first

  1. 2024M3oE: Multi-Domain Multi-Task Mixture-of Experts Recommendation Framework29 Apr

Other references · 21

  1. 2017FiLM: Visual Reasoning with a General Conditioning Layercited by 2,286
  2. 2017GradNorm: Gradient Normalization for Adaptive Loss Balancing in Deep Multitask Networkscited by 816
  3. 2022A Review of Sparse Expert Models in Deep Learningcited by 147
  4. 2022KuaiRec: A Fully-observed Dataset and Insights for Evaluating Recommender Systemscited by 94
  5. 2024Merging Multi-Task Models via Weight-Ensembling Mixture of Expertscited by 85
  6. 2021One Model to Serve All: Star Topology Adaptive Recommender for Multi-Domain CTR Predictioncited by 78
  7. 2019GNN-FiLM: Graph Neural Networks with Feature-wise Linear Modulationcited by 75
  8. 2023PEPNet: Parameter and Embedding Personalized Network for Infusing with Personalized Prior Informationcited by 60
  9. 2019DEAR: Deep Reinforcement Learning for Online Advertising Impression in Recommender Systemscited by 44
  10. 2019The Fairness of Risk Scores Beyond Classification: Bipartite Ranking and the xAUC Metriccited by 41
  11. 2022Tenrec: A Large-scale Multipurpose Benchmark Dataset for Recommender Systemscited by 31
  12. 2020Interplay between Upsampling and Regularization for Provider Fairness in Recommender Systemscited by 15
9 more
  1. 2019Recommender Systems with Heterogeneous Side Informationcited by 11
  2. 2018Personalizing Fairness-aware Re-rankingcited by 9
  3. 2021Heterogeneous Graph Neural Networks for Large-Scale Bid Keyword Matchingcited by 6
  4. 2025Preserving AUC Fairness in Learning with Noisy Protected Groupscited by 6
  5. 2024Retrievable Domain-Sensitive Feature Memory for Multi-Domain Recommendationcited by 4
  6. 2022Deep Unified Representation for Heterogeneous Recommendationcited by 3
  7. 2025Representational Alignment Across Model Layers and Brain Regions with Multi-Level Optimal Transportcited by 3
  8. 2023Distributionally-Informed Recommender System Evaluationcited by 2
  9. 2024A Parameter Update Balancing Algorithm for Multi-task Ranking Models in Recommendation Systemscited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025MTmixAtt: Integrating Mixture-of-Experts with Multi-Mix Attention for Large-Scale Recommendation
  2. 2018Recommendation Through Mixtures of Heterogeneous Item Relationships
  3. 2026SMES: Towards Scalable Multi-Task Recommendation via Expert Sparsity
  4. 2025RankMixer: Scaling Up Ranking Models in Industrial Recommenders
  5. 2026Generate to Explore, Select to Exploit: Aligning LLM-based Headline Generation with Personalized Recommendation
  6. 2026OneTrans-V2: Unifying Retrieval, Pre-rank, and Fine-rank with One Transformer in Industrial Recommender
  7. 2026X-KGRank: A Knowledge Graph RAG Framework for Explainable Recommendations via Pattern Mining and LLM Re-Ranking
  8. 2025A Survey of Foundation Model-Powered Recommender Systems: From Feature-Based, Generative to Agentic Paradigms