aiqbench analysis
Meta has deployed a new "Foundation-Expert" paradigm for recommender systems that addresses the challenge of scaling foundation models across multiple platforms. By using target-aware embeddings as input features rather than soft labels, this approach significantly improves transfer fidelity and delivers measurable gains on billions of daily requests. This work represents a concrete advancement in applying large-scale AI architectures to high-volume recommendation tasks at hyperscale.
Realizing Scaling Laws in Recommender Systems: A Foundation-Expert Paradigm for Hyperscale Model Deployment
arXiv:2508.02929v3 Announce Type: replace-cross 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.
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.
Information Retrieval (cs.IR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
arXiv:2508.02929 [cs.IR]
(or arXiv:2508.02929v3 [cs.IR] for this version)
Full text under the publisher's open licence · 188 words
Related coverage
Jev: System One models for Prod, not God — with Diogo Almeida, CEO, TypeSafe AI
Latent Space · Sep 21, 2026
Dyna-2: General Purpose Physical Agents with Hour-Long Autonomy — Jason Ma, Dyna Robotics
Latent Space · Oct 13, 2026
Links #5: 2026/07
LessWrong · Jul 31, 2026
Game Design Creativity evals
LessWrong · Oct 12, 2026