External Large Foundation Model: How to Efficiently Serve Trillions of Parameters for Online Ads Recommendation

TL;DR

This paper introduces ExFM, a framework for efficiently serving trillion-parameter models in online ads recommendation, addressing challenges of computational cost and dynamic data distribution with novel distillation and data augmentation techniques.

Contribution

The paper proposes the External Large Foundation Model (ExFM) framework, including external distillation, data augmentation, and a multi-tenant teacher-student design for scalable, efficient online recommendation.

Findings

Significant performance improvements on industrial-scale applications.

Effective control of training and inference costs.

Robustness to dynamic data distribution shifts.

Abstract

Ads recommendation is a prominent service of online advertising systems and has been actively studied. Recent studies indicate that scaling-up and advanced design of the recommendation model can bring significant performance improvement. However, with a larger model scale, such prior studies have a significantly increasing gap from industry as they often neglect two fundamental challenges in industrial-scale applications. First, training and inference budgets are restricted for the model to be served, exceeding which may incur latency and impair user experience. Second, large-volume data arrive in a streaming mode with data distributions dynamically shifting, as new users/ads join and existing users/ads leave the system. We propose the External Large Foundation Model (ExFM) framework to address the overlooked challenges. Specifically, we develop external distillation and a data augmentation system (DAS) to control the computational cost of training/inference while maintaining high performance. We design the teacher in a way like a foundation model (FM) that can serve multiple students as vertical models (VMs) to amortize its building cost. We propose Auxiliary Head and Student Adapter to mitigate the data distribution gap between FM and VMs caused by the streaming data issue. Comprehensive experiments on internal industrial-scale applications and public datasets demonstrate significant performance gain by ExFM.