DisCo: Graph-Based Disentangled Contrastive Learning for Cold-Start Cross-Domain Recommendation

TL;DR

DisCo introduces a graph-based disentangled contrastive learning framework for cross-domain recommendation that effectively captures user intents and filters irrelevant information to improve cold-start user predictions.

Contribution

It proposes a novel graph-based disentangled contrastive learning method that enhances cross-domain recommendation by capturing fine-grained user intents and reducing negative transfer.

Findings

DisCo outperforms state-of-the-art baselines on four benchmark datasets.

The framework effectively captures diverse user intents.

DisCo reduces negative transfer in cold-start scenarios.

Abstract

Recommender systems are widely used in various real-world applications, but they often encounter the persistent challenge of the user cold-start problem. Cross-domain recommendation (CDR), which leverages user interactions from one domain to improve prediction performance in another, has emerged as a promising solution. However, users with similar preferences in the source domain may exhibit different interests in the target domain. Therefore, directly transferring embeddings may introduce irrelevant source-domain collaborative information. In this paper, we propose a novel graph-based disentangled contrastive learning framework to capture fine-grained user intent and filter out irrelevant collaborative information, thereby avoiding negative transfer. Specifically, for each domain, we use a multi-channel graph encoder to capture diverse user intents. We then construct the affinity graph in the embedding space and perform multi-step random walks to capture high-order user similarity relationships. Treating one domain as the target, we propose a disentangled intent-wise contrastive learning approach, guided by user similarity, to refine the bridging of user intents across domains. Extensive experiments on four benchmark CDR datasets demonstrate that DisCo consistently outperforms existing state-of-the-art baselines, thereby validating the effectiveness of both DisCo and its components.