MDiffFR: Modality-Guided Diffusion Generation for Cold-start Items in Federated Recommendation

TL;DR

MDiffFR introduces a diffusion-based approach guided by modality features to generate embeddings for cold-start items in federated recommendation systems, effectively addressing data privacy constraints and improving recommendation accuracy.

Contribution

The paper proposes a novel diffusion model guided by modality features for cold-start item embedding generation in federated recommendation systems, overcoming limitations of previous attribute-to-embedding mappings.

Findings

Outperforms baseline methods on four real datasets

Provides stronger privacy guarantees than existing approaches

Successfully generates effective embeddings for cold-start items

Abstract

Federated recommendations (FRs) provide personalized services while preserving user privacy by keeping user data on local clients, which has attracted significant attention in recent years. However, due to the strict privacy constraints inherent in FRs, access to user-item interaction data and user profiles across clients is highly restricted, making it difficult to learn globally effective representations for new (cold-start) items. Consequently, the item cold-start problem becomes even more challenging in FRs. Existing solutions typically predict embeddings for new items through the attribute-to-embedding mapping paradigm, which establishes a fixed one-to-one correspondence between item attributes and their embeddings. However, this one-to-one mapping paradigm often fails to model varying data distributions and tends to cause embedding misalignment, as verified by our empirical studies. To this end, we propose MDiffFR, a novel generation-based modality-guided diffusion method for cold-start items in FRs. In this framework, we employ a tailored diffusion model on the server to generate embeddings for new items, which are then distributed to clients for cold-start inference. To align item semantics, we deploy a pre-trained modality encoder to extract modality features as conditional signals to guide the reverse denoising process. Furthermore, our theoretical analysis verifies that the proposed method achieves stronger privacy guarantees compared to existing mapping-based approaches. Extensive experiments on four real datasets demonstrate that our method consistently outperforms all baselines in FRs.