{"ID":2877921,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2508.18700","arxiv_id":"2508.18700","title":"Taming the One-Epoch Phenomenon in Online Recommendation System by Two-stage Contrastive ID Pre-training","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.","short_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...","url_abs":"https://arxiv.org/abs/2508.18700","url_pdf":"https://arxiv.org/pdf/2508.18700v1","authors":"[\"Yi-Ping Hsu\",\"Po-Wei Wang\",\"Chantat Eksombatchai\",\"Jiajing Xu\"]","published":"2025-08-26T06:06:21Z","proceeding":"cs.IR","tasks":"[\"cs.IR\",\"cs.LG\"]","methods":"[]","has_code":false}