mlsanityData to 14 Oct 2026

Scaling User Modeling: Large-scale Online User Representations for Ads Personalization in Meta

arXiv 2311.09544 ↗First posted 16 Nov 2023cs.IR cs.AI cs.LGcited by 21

Not on the map. Why?

Abstract (arXiv)

Effective user representations are pivotal in personalized advertising. However, stringent constraints on training throughput, serving latency, and memory, often limit the complexity and input feature set of online ads ranking models. This challenge is magnified in extensive systems like Meta's, which encompass hundreds of models with diverse specifications, rendering the tailoring of user representation learning for each model impractical. To address these challenges, we present Scaling User Modeling (SUM), a framework widely deployed in Meta's ads ranking system, designed to facilitate efficient and scalable sharing of online user representation across hundreds of ads models. SUM leverages a few designated upstream user models to synthesize user embeddings from massive amounts of user features with advanced modeling techniques. These embeddings then serve as inputs to downstream online ads ranking models, promoting efficient representation sharing. To adapt to the dynamic nature of user features and ensure embedding freshness, we designed SUM Online Asynchronous Platform (SOAP), a latency free online serving system complemented with model freshness and embedding stabilization, which enables frequent user model updates and online inference of user embeddings upon each user request. We share our hands-on deployment experiences for the SUM framework and validate its superiority through comprehensive experiments. To date, SUM has been launched to hundreds of ads ranking models in Meta, processing hundreds of billions of user requests daily, yielding significant online metric gains and improved infrastructure efficiency.

Key facts. Cited by 21 papers in the map (data through 14 Oct 2026). Builds on 23 of its 36 references that are in the map. First posted on arXiv 16 Nov 2023.

Builds on

23 of 36 references.

References · 23

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2021MLP-Mixer: An all-MLP Architecture for Visioncited by 1,446
  3. 2018Graph Convolutional Neural Networks for Web-Scale Recommender Systemscited by 1,138
  4. 2017Neural Collaborative Filteringcited by 1,111
  5. 2016Wide & Deep Learning for Recommender Systemscited by 747
  6. 2017DeepFM: A Factorization-Machine based Neural Network for CTR Predictioncited by 676
  7. 2017Deep & Cross Network for Ad Click Predictionscited by 393
  8. 2018Deep Interest Evolution Network for Click-Through Rate Predictioncited by 338
  9. 2019Deep Learning Recommendation Model for Personalization and Recommendation Systemscited by 327
  10. 2020DCN V2: Improved Deep & Cross Network and Practical Lessons for Web-scale Learning to Rank Systemscited by 327
  11. 2019Practice on Long Sequential User Behavior Modeling for Click-Through Rate Predictioncited by 118
  12. 2018Billion-scale Commodity Embedding for E-commerce Recommendation in Alibabacited by 105
11 more
  1. 2017Personalizing Session-based Recommendations with Hierarchical Recurrent Neural Networkscited by 84
  2. 2022PinnerFormer: Sequence Modeling for User Representation at Pinterestcited by 57
  3. 2013Empirical Analysis of Predictive Algorithms for Collaborative Filteringcited by 47
  4. 2023TransAct: Transformer-based Realtime User Action Model for Recommendation at Pinterestcited by 35
  5. 2020One Person, One Model, One World: Learning Continual User Representation without Forgettingcited by 30
  6. 2022TwHIN: Embedding the Twitter Heterogeneous Information Network for Personalized Recommendationcited by 25
  7. 2022ItemSage: Learning Product Embeddings for Shopping Recommendations at Pinterestcited by 23
  8. 2022AutoShard: Automated Embedding Table Sharding for Recommender Systemscited by 17
  9. 2020General-Purpose User Embeddings based on Mobile App Usagecited by 11
  10. 2023On the Consistency of Average Embeddings for Item Recommendationcited by 5
  11. 2019A Feature Selection Based on Perturbation Theorycited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 3; 2025: 9; 2026: 9.

21 papers.

Sort:

By area:

  1. 2025External Large Foundation Model: How to Efficiently Serve Trillions of Parameters for Online Ads Recommendationcited by 13
  2. 2025You Are What You Bought: Generating Customer Personas for E-commerce Applicationscited by 7
  3. 2026LLaTTE: Scaling Laws for Multi-Stage Sequence Modeling in Large-Scale Ads Recommendationcited by 5
  4. 2025Realizing Scaling Laws in Recommender Systems: A Foundation-Expert Paradigm for Hyperscale Model Deploymentcited by 4
  5. 2025FinTRec: Transformer Based Unified Contextual Ads Targeting and Personalization for Financial Applicationscited by 4
  6. 2025DashCLIP: Leveraging multimodal models for generating semantic embeddings for DoorDashcited by 2
  7. 2025RecIS: Sparse to Dense, A Unified Training Framework for Recommendation Modelscited by 2
  8. 2024ERCache: An Efficient and Reliable Caching Framework for Large-Scale User Representations in Meta's Ads Systemcited by 1
  9. 2025Instruction-aware User Embedding via Synergistic Language and Representation Modelingcited by 1
  10. 2026Query as Anchor: Scenario-Adaptive User Representation via Large Language Modelcited by 1
  11. 2026LoKA: Low-precision Kernel Applications for Recommendation Models At Scalecited by 1
  12. 2024Pfeed: Generating near real-time personalized feeds using precomputed embedding similaritiescited by 0
9 more
  1. 2024Root Causing Prediction Anomalies Using Explainable AIcited by 0
  2. 2025Large Foundation Model for Ads Recommendationcited by 0
  3. 2025Decoupled Entity Representation Learning for Pinterest Ads Rankingcited by 0
  4. 2026LoopFM: Learning frOm HistOrical RePresentations of Foundation Model for Recommendationcited by 0
  5. 2026DUET -- Dual User Embedding Transformers for Offsite Conversion Predictioncited by 0
  6. 2026Don't Contrast the Impossible: Region-Constrained Batching for Contrastive User Modeling on a Local Community Platformcited by 0
  7. 2026PersonaMem-v3: Toward Omni-Platform Personal Intelligence for Holistic User Understanding, Recommendation, and Agentic Taskscited by 0
  8. 2026DP-Rec: Towards Dynamic Patching for Efficient Long-Sequence Recommendationcited by 0
  9. 2026CUEing User Simulators: Calibrated User Embeddings for Multi-Turn Benchmarkingcited by 0