Diffusion Cross-domain Recommendation

TL;DR

This paper addresses cold-start challenges in recommender systems by proposing DiffCDR, a diffusion-based cross-domain recommendation framework that transfers knowledge from a rich source domain to a sparse target domain. The model combines a Diffusion Module (DIM) that generates target-domain user embeddings conditioned on source-domain information with an Alignment Module (ALM) to stabilize transfers, plus a task-oriented loss to tailor optimization to ranking and rating performance. Across three Amazon-based CDR tasks, DiffCDR demonstrates improved cold-start and warm-start performance, with ablations showing that each component (DIM, ALM, and task loss) contributes to gains. The approach also highlights practical aspects such as inference speed through fast diffusion solvers and reasonable throughput, suggesting real-world applicability for cross-domain recommendations.

Abstract

It is always a challenge for recommender systems to give high-quality outcomes to cold-start users. One potential solution to alleviate the data sparsity problem for cold-start users in the target domain is to add data from the auxiliary domain. Finding a proper way to extract knowledge from an auxiliary domain and transfer it into a target domain is one of the main objectives for cross-domain recommendation (CDR) research. Among the existing methods, mapping approach is a popular one to implement cross-domain recommendation models (CDRs). For models of this type, a mapping module plays the role of transforming data from one domain to another. It primarily determines the performance of mapping approach CDRs. Recently, diffusion probability models (DPMs) have achieved impressive success for image synthesis related tasks. They involve recovering images from noise-added samples, which can be viewed as a data transformation process with outstanding performance. To further enhance the performance of CDRs, we first reveal the potential connection between DPMs and mapping modules of CDRs, and then propose a novel CDR model named Diffusion Cross-domain Recommendation (DiffCDR). More specifically, we first adopt the theory of DPM and design a Diffusion Module (DIM), which generates user's embedding in target domain. To reduce the negative impact of randomness introduced in DIM and improve the stability, we employ an Alignment Module to produce the aligned user embeddings. In addition, we consider the label data of the target domain and form the task-oriented loss function, which enables our DiffCDR to adapt to specific tasks. By conducting extensive experiments on datasets collected from reality, we demonstrate the effectiveness and adaptability of DiffCDR to outperform baseline models on various CDR tasks in both cold-start and warm-start scenarios.

Figures (4)

• Figure 1: Illustration of the DiffCDR concept. (a) Learning the Diffusion Module (DIM) $p_\theta$ on the embeddings of overlapping users (b) Generating transferred embeddings via reverse diffusion process for cold-start users conditioned on corresponding latent vectors from auxiliary domain.
• Figure 2: Architecture of DiffCDR. It consists of a Diffusion Module (DIM) and an Alignment Module (ALM). We also employ a diffusion solver to accelerate the inference of the DIM. The recommendation results are produced by element-wisely multiplication with the output of ALM and the item embeddings in target domain.
• Figure 3: Results of warm-start experiments. The light and dark colored represent the cold-start and warm-start respectively.
• Figure 4: User embedding visualization. We randomly sample 1000 users and visualize embeddings of these users in both train and test stages. (a,b) involve different CDRs, and (c,d) are produced by DiffCDR with various inference steps.