ID-centric Pre-training for Recommendation

TL;DR

This paper introduces an ID-centric pre-training approach for recommendation systems that transfers ID embeddings across domains, leveraging a cross-domain ID-matcher to improve recommendation accuracy especially in cold-start scenarios.

Contribution

The work proposes a novel ID-centric pre-training paradigm (IDP) with a cross-domain ID-matcher (CDIM) that directly transfers ID embeddings across domains, enhancing recommendation performance.

Findings

Significantly outperforms baseline models in experiments.

Effective in both cold and warm start settings.

Utilizes textual information to retrieve similar pre-trained ID embeddings.

Abstract

Classical sequential recommendation models generally adopt ID embeddings to store knowledge learned from user historical behaviors and represent items. However, these unique IDs are challenging to be transferred to new domains. With the thriving of pre-trained language model (PLM), some pioneer works adopt PLM for pre-trained recommendation, where modality information (e.g., text) is considered universal across domains via PLM. Unfortunately, the behavioral information in ID embeddings is still verified to be dominating in PLM-based recommendation models compared to modality information and thus limits these models' performance. In this work, we propose a novel ID-centric recommendation pre-training paradigm (IDP), which directly transfers informative ID embeddings learned in pre-training domains to item representations in new domains. Specifically, in pre-training stage, besides the ID-based sequential model for recommendation, we also build a Cross-domain ID-matcher (CDIM) learned by both behavioral and modality information. In the tuning stage, modality information of new domain items is regarded as a cross-domain bridge built by CDIM. We first leverage the textual information of downstream domain items to retrieve behaviorally and semantically similar items from pre-training domains using CDIM. Next, these retrieved pre-trained ID embeddings, rather than certain textual embeddings, are directly adopted to generate downstream new items' embeddings. Through extensive experiments on real-world datasets, both in cold and warm settings, we demonstrate that our proposed model significantly outperforms all baselines. Codes will be released upon acceptance.