REVIEW 5 cited by
Realizing Scaling Laws in Recommender Systems: A Foundation-Expert Paradigm for Hyperscale Model Deployment
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Realizing Scaling Laws in Recommender Systems: A Foundation-Expert Paradigm for Hyperscale Model Deployment
read the original abstract
Scaling laws have been established for recommender systems, yet efficiently deploying foundation model (FM) across multiple recommendation surfaces remains a major unsolved challenge. Existing methods for transfer learning face fundamental limitations in this setting: knowledge distillation suffers from transfer fidelity degradation in the large-data regime, and static user or item embeddings lack the expressiveness to capture contextualized user-item interactions. We propose the Foundation-Expert paradigm, where a central FM generates target-aware embeddings which are ingested by lightweight surface-specific expert models. Target-aware embeddings are representations that dynamically capture a user's interest in a specific item conditioned on their full interaction history. Unlike knowledge distillation, which transfers FM knowledge as soft labels, our approach provides these embeddings as input features to each expert model, enabling direct interaction with surface-specific representations. This paradigm achieves transfer ratios of 0.64--1.0 from FM to experts, substantially exceeding existing methods. Fully deployed at Meta serving tens of billions of daily requests since 2025, it delivers 0.050% statistically significant online topline metric improvement and 0.359% cumulative gains across multiple surfaces.
Forward citations
Cited by 5 Pith papers
-
OneShot: Index-in-Ranking with Neural Scoring for Large-Scale Retrieval
OneShot trains hierarchical item codebooks jointly with the ranking loss, enabling nonlinear neural scoring in billion-scale retrieval and reporting +20% recall, 10x fewer dense-ranked items, and live Instagram gains.
-
OneShot: Index-in-Ranking with Neural Scoring for Large-Scale Retrieval
OneShot trains a hierarchical item index and a neural ranking score jointly, reporting +20% offline recall and 10x efficiency over k-means ANN at Instagram.
-
Mosaic: A Fleet of User Embedding Specialists for Recommendation at Meta
Mosaic shows that a fleet of four heterogeneous user-embedding specialists, trained with redundancy-reduction and composite-label losses, improves downstream recommendation quality at Meta.
-
Versioned Late Materialization for Ultra-Long Sequence Training in Recommendation Systems at Scale
Versioned late materialization stores user histories once and reconstructs sequences just-in-time during training to cut redundancy and enable longer sequences in large-scale recommendation systems.
-
Versioned Late Materialization for Ultra-Long Sequence Training in Recommendation Systems at Scale
Introduces versioned late materialization to eliminate data redundancy in ultra-long sequence training for DLRMs by storing histories once and reconstructing via pointers at training time.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.