LiMAML: Personalization of Deep Recommender Models via Meta Learning

TL;DR

LiMAML addresses the need for fast, per-entity personalization in large-scale recommender systems by introducing a two-block architecture where a Meta Block is meta-learned to produce fixed-size meta embeddings, which condition a shared Global Block during online serving. By training only the meta block in the inner loop and storing per-entity embeddings rather than full per-entity model weights, LiMAML achieves personalization at scale with production-friendly storage and latency characteristics. Extensive offline and online experiments across LinkedIn applications show consistent improvements in AUC and CTR over baselines and strong baselines like wide-and-deep ID embeddings, with notable gains on low-data or infrequent users. The results demonstrate the practicality and impact of meta-learning-based personalization for real-world, high-volume recommender systems, and outline promising directions for integrating with larger foundation models and multi-entity personalization strategies.

Abstract

In the realm of recommender systems, the ubiquitous adoption of deep neural networks has emerged as a dominant paradigm for modeling diverse business objectives. As user bases continue to expand, the necessity of personalization and frequent model updates have assumed paramount significance to ensure the delivery of relevant and refreshed experiences to a diverse array of members. In this work, we introduce an innovative meta-learning solution tailored to the personalization of models for individual members and other entities, coupled with the frequent updates based on the latest user interaction signals. Specifically, we leverage the Model-Agnostic Meta Learning (MAML) algorithm to adapt per-task sub-networks using recent user interaction data. Given the near infeasibility of productionizing original MAML-based models in online recommendation systems, we propose an efficient strategy to operationalize meta-learned sub-networks in production, which involves transforming them into fixed-sized vectors, termed meta embeddings, thereby enabling the seamless deployment of models with hundreds of billions of parameters for online serving. Through extensive experimentation on production data drawn from various applications at LinkedIn, we demonstrate that the proposed solution consistently outperforms the baseline models of those applications, including strong baselines such as using wide-and-deep ID based personalization approach. Our approach has enabled the deployment of a range of highly personalized AI models across diverse LinkedIn applications, leading to substantial improvements in business metrics as well as refreshed experience for our members.

Figures (1)

• Figure 1: LiMAML Model Structure