Taming the One-Epoch Phenomenon in Online Recommendation System by Two-stage Contrastive ID Pre-training

TL;DR

This paper proposes a two-stage contrastive pre-training approach for ID-based embeddings in online recommendation systems, addressing the one-epoch overfitting problem and improving online performance.

Contribution

It introduces a novel two-stage training strategy with contrastive pre-training that enhances data coverage and online generalization in recommendation embeddings.

Findings

Multi-epoch pre-training does not cause overfitting.

Embeddings from pre-training improve online recommendation performance.

Deployment at Pinterest led to significant engagement gains.

Abstract

ID-based embeddings are widely used in web-scale online recommendation systems. However, their susceptibility to overfitting, particularly due to the long-tail nature of data distributions, often limits training to a single epoch, a phenomenon known as the "one-epoch problem." This challenge has driven research efforts to optimize performance within the first epoch by enhancing convergence speed or feature sparsity. In this study, we introduce a novel two-stage training strategy that incorporates a pre-training phase using a minimal model with contrastive loss, enabling broader data coverage for the embedding system. Our offline experiments demonstrate that multi-epoch training during the pre-training phase does not lead to overfitting, and the resulting embeddings improve online generalization when fine-tuned for more complex downstream recommendation tasks. We deployed the proposed system in live traffic at Pinterest, achieving significant site-wide engagement gains.