mlsanityData to 14 Oct 2026

Pixie: A System for Recommending 3+ Billion Items to 200+ Million Users in Real-Time

arXiv 1711.07601 ↗First posted 21 Nov 2017cs.IR cs.LG cs.PFcited by 47

Not on the map. Why?

Abstract (arXiv)

User experience in modern content discovery applications critically depends on high-quality personalized recommendations. However, building systems that provide such recommendations presents a major challenge due to a massive pool of items, a large number of users, and requirements for recommendations to be responsive to user actions and generated on demand in real-time. Here we present Pixie, a scalable graph-based real-time recommender system that we developed and deployed at Pinterest. Given a set of user-specific pins as a query, Pixie selects in real-time from billions of possible pins those that are most related to the query. To generate recommendations, we develop Pixie Random Walk algorithm that utilizes the Pinterest object graph of 3 billion nodes and 17 billion edges. Experiments show that recommendations provided by Pixie lead up to 50% higher user engagement when compared to the previous Hadoop-based production system. Furthermore, we develop a graph pruning strategy at that leads to an additional 58% improvement in recommendations. Last, we discuss system aspects of Pixie, where a single server executes 1,200 recommendation requests per second with 60 millisecond latency. Today, systems backed by Pixie contribute to more than 80% of all user engagement on Pinterest.

Key facts. Cited by 47 papers in the map (data through 14 Oct 2026). First posted on arXiv 21 Nov 2017.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 8; 2022: 10; 2023: 9; 2024: 6; 2025: 7; 2026: 7.

47 papers.

Sort:

By area:

  1. 2023Evaluating Graph Neural Networks for Link Prediction: Current Pitfalls and New Benchmarkingcited by 69
  2. 2023Leveraging Large Language Models in Conversational Recommender Systemscited by 63
  3. 2021Sampling methods for efficient training of graph convolutional networks: A surveycited by 40
  4. 2022ItemSage: Learning Product Embeddings for Shopping Recommendations at Pinterestcited by 23
  5. 2021Persia: An Open, Hybrid System Scaling Deep Learning-based Recommenders up to 100 Trillion Parameterscited by 17
  6. 2022Rethinking Personalized Ranking at Pinterest: An End-to-End Approachcited by 16
  7. 2021On component interactions in two-stage recommender systemscited by 14
  8. 2023Revisiting Neural Retrieval on Acceleratorscited by 14
  9. 2024HiLo: Detailed and Robust 3D Clothed Human Reconstruction with High-and Low-Frequency Information of Parametric Modelscited by 13
  10. 2021Blockchain-based Recommender Systems: Applications, Challenges and Future Opportunitiescited by 11
  11. 2022Uncertainty Quantification for Fairness in Two-Stage Recommender Systemscited by 10
  12. 2024Machine Unlearning for Recommendation Systems: An Insightcited by 9
35 more
  1. 2021Local Hyper-Flow Diffusioncited by 8
  2. 2021GraphTheta: A Distributed Graph Neural Network Learning System With Flexible Training Strategycited by 8
  3. 2022MultiBiSage: A Web-Scale Recommendation System Using Multiple Bipartite Graphs at Pinterestcited by 6
  4. 2022ZOOMER: Boosting Retrieval on Web-scale Graphs by Regions of Interestcited by 4
  5. 2024Graph Feature Preprocessor: Real-time Subgraph-based Feature Extraction for Financial Crime Detectioncited by 4
  6. 2022Distributed Graph Neural Network Training with Periodic Stale Representation Synchronizationcited by 3
  7. 2024Retrieval with Learned Similaritiescited by 3
  8. 2022A Design of A Simple Yet Effective Exercise Recommendation System in K-12 Online Learningcited by 2
  9. 2023Local Graph Clustering with Noisy Labelscited by 2
  10. 2025A Two-Stage Data Selection Framework for Data-Efficient Model Training on Edge Devicescited by 2
  11. 2026RelayGR: Scaling Long-Sequence Generative Recommendation via Cross-Stage Relay-Race Inferencecited by 2
  12. 2021Frame invariance and scalability of neural operators for partial differential equationscited by 1
  13. 2022Hardware/Software Co-Programmable Framework for Computational SSDs to Accelerate Deep Learning Service on Large-Scale Graphscited by 1
  14. 2022Evolution of a Web-Scale Near Duplicate Image Detection Systemcited by 1
  15. 2023Efficient Large-scale Nonstationary Spatial Covariance Function Estimation Using Convolutional Neural Networkscited by 1
  16. 2023EHI: End-to-end Learning of Hierarchical Index for Efficient Dense Retrievalcited by 1
  17. 2025Towards Principled Learning for Re-ranking in Recommender Systemscited by 1
  18. 2025Privacy Preserving Inference of Personalized Content for Out of Matrix Userscited by 1
  19. 2025Near-Zero-Overhead Freshness for Recommendation Systems via Inference-Side Model Updatescited by 1
  20. 2026PinEqualizer: Full Funnel Content Exploration and Debiasing System at Pinterestcited by 1
  21. 2021Thematic recommendations on knowledge graphs using multilayer networkscited by 0
  22. 2022GraTO: Graph Neural Network Framework Tackling Over-smoothing with Neural Architecture Searchcited by 0
  23. 2023A Dataset for Learning Graph Representations to Predict Customer Returns in Fashion Retailcited by 0
  24. 2023Decision Support System for Chronic Diseases Based on Drug-Drug Interactionscited by 0
  25. 2023Graph Rewriting for Graph Neural Networkscited by 0
  26. 2024Towards a Theoretical Understanding of Two Tower Recommendation Modelscited by 0
  27. 2024Modeling Social Media Recommendation Impacts Using Academic Networks: A Graph Neural Network Approachcited by 0
  28. 2025GPU-accelerated Multi-relational Parallel Graph Retrieval for Web-scale Recommendationscited by 0
  29. 2025Creator-Side Recommender System: Challenges, Designs, and Applicationscited by 0
  30. 2025Graph Neural Networks, Deep Reinforcement Learning and Probabilistic Topic Modeling for Strategic Multiagent Settingscited by 0
  31. 2026PinCLIP: Large-scale Foundational Multimodal Representation at Pinterestcited by 0
  32. 2026FragBench: Cross-Session Attacks Hidden in Benign-Looking Fragmentscited by 0
  33. 2026Keeping the Index Open: The Recommendation-Side Cost of Shared Search and Recommendationcited by 0
  34. 2026A Multi-Source Ensemble Approach to Candidate Generation for Alternative Vacation Rental Property Recommendationscited by 0
  35. 2026RT-J: Large-Scale Pretraining of Relational Transformers for Context-Efficient Predictionscited by 0