mlsanityData to 14 Oct 2026

DV365: Extremely Long User History Modeling at Instagram

arXiv 2506.00450 ↗First posted 31 May 2025cs.IR cs.LGcited by 8

Not on the map. Why?

Abstract (arXiv)

Long user history is highly valuable signal for recommendation systems, but effectively incorporating it often comes with high cost in terms of data center power consumption and GPU. In this work, we chose offline embedding over end-to-end sequence length optimization methods to enable extremely long user sequence modeling as a cost-effective solution, and propose a new user embedding learning strategy, multi-slicing and summarization, that generates highly generalizable user representation of user's long-term stable interest. History length we encoded in this embedding is up to 70,000 and on average 40,000. This embedding, named as DV365, is proven highly incremental on top of advanced attentive user sequence models deployed in Instagram. Produced by a single upstream foundational model, it is launched in 15 different models across Instagram and Threads with significant impact, and has been production battle-proven for >1 year since our first launch.

Key facts. Cited by 8 papers in the map (data through 14 Oct 2026). Builds on 13 of its 25 references that are in the map. First posted on arXiv 31 May 2025.

Builds on

13 of 25 references.

References · 13

  1. 2018Self-Attentive Sequential Recommendationcited by 1,184
  2. 2017Deep Interest Network for Click-Through Rate Predictioncited by 544
  3. 2018Deep Interest Evolution Network for Click-Through Rate Predictioncited by 338
  4. 2020Search-based User Interest Modeling with Lifelong Sequential Behavior Data for Click-Through Rate Predictioncited by 177
  5. 2020Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processingcited by 162
  6. 2019Practice on Long Sequential User Behavior Modeling for Click-Through Rate Predictioncited by 118
  7. 2021Software-Hardware Co-design for Fast and Scalable Training of Deep Learning Recommendation Modelscited by 78
  8. 2022PinnerFormer: Sequence Modeling for User Representation at Pinterestcited by 57
  9. 2019Lifelong Sequential Modeling with Personalized Memorization for User Response Predictioncited by 41
  10. 2022Sampling Is All You Need on Modeling Long-Term User Behaviors for CTR Predictioncited by 24
  11. 2023Revisiting Neural Retrieval on Acceleratorscited by 14
  12. 2020GateNet: Gating-Enhanced Deep Network for Click-Through Rate Predictioncited by 10
1 more
  1. 2021Empowering General-purpose User Representation with Full-life Cycle Behavior Modelingcited by 6

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 8.

8 papers.

  1. 2026High Fidelity Textual User Representation over Heterogeneous Sources via Reinforcement Learningcited by 1
  2. 2026Fine-Tuned LLM as a Complementary Predictor Improving Ads Systemcited by 1
  3. 2026TokenMinds: Pretrained User Tokens and Embeddings for User Understanding in Large Recommender Systemscited by 1
  4. 2026Efficient Dataset Selection for Continual Adaptation of Generative Recommenderscited by 0
  5. 2026Versioned Late Materialization for Ultra-Long Sequence Training in Recommendation Systems at Scalecited by 0
  6. 2026UxSID: Semantic-Aware User Interests Modeling for Ultra-Long Sequencecited by 0
  7. 2026Beyond Item IDs: Scaling Short-Form-Video Recommendation via Semantic-Native Long Sequence Modelingcited by 0
  8. 2026A Foundation Model for Multimodal Event Sequences in Financial Applicationscited by 0