The author argues that a Privacy-Preserving Framework with Multi-Modal Data for Cross-Domain Recommendation, called P2M2-CDR, effectively addresses challenges in disentangling domain-common and domain-specific features while protecting user privacy during knowledge transfer.
The core message of this paper is that the distortion of user similarity relationships across domains is a key factor causing negative transfer in cross-domain recommendation, and the proposed Collaborative information regularized User Transformation (CUT) framework can effectively alleviate this issue by directly filtering irrelevant source-domain collaborative information.
A unified framework, UniCDR+, is proposed to effectively transfer knowledge across multiple domains and adapt to various cross-domain recommendation scenarios, outperforming existing task-specific approaches.
The core message of this paper is that not all information from the source domain is equally beneficial for cross-domain recommendation (CDR) tasks. The authors propose a novel framework, CoTrans, that selectively compresses and transfers relevant knowledge from the source domain to the target domain, guided by the target domain's information.
DIIT, a novel method for industrial cross-domain recommendation, leverages domain-invariant information from multiple source domains to enhance recommendation effectiveness in a target domain, addressing the limitations of traditional methods in industrial recommender systems.
This paper proposes HAGO, a novel data-centric framework that leverages heterogeneous adaptive graph coordinators and graph prompting to address the challenges of cross-domain recommendation by aligning representations and transferring knowledge across multiple domains.