Federated Semantic Learning for Privacy-preserving Cross-domain Recommendation

2025-03-29

AI Summary

Addressing the dual challenges of privacy preservation and semantic modeling in non-overlapping cross-domain recommendation (CDR), this paper proposes FFMSR, a Federated Framework for Multi-modal Semantic Recommendation. Existing approaches suffer from textual semantic loss, unimodal modeling limitations, and interference from irrelevant features. FFMSR makes three key contributions: (1) a hierarchical semantic encoder that preserves rich lexical semantics without distortion; (2) an ID-Text dual-modality collaborative modeling mechanism to enhance user representation robustness; and (3) a frequency-domain filtering module based on Fast Fourier Transform (FFT) combined with a gating mechanism to suppress irrelevant semantic noise. On the server side, semantic clustering enables cross-domain knowledge transfer, while end-to-end federated learning guarantees data privacy. Extensive experiments on two real-world datasets demonstrate significant improvements over state-of-the-art methods. The source code is publicly available.

Technology Category

• Machine Learning: Multimodal Learning
• Application Domains: Humanities & Computational Social Science
• Computer Vision: Multi-modal Vision

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

In the evolving landscape of recommender systems, the challenge of effectively conducting privacy-preserving Cross-Domain Recommendation (CDR), especially under strict non-overlapping constraints, has emerged as a key focus. Despite extensive research has made significant progress, several limitations still exist: 1) Previous semantic-based methods fail to deeply exploit rich textual information, since they quantize the text into codes, losing its original rich semantics. 2) The current solution solely relies on the text-modality, while the synergistic effects with the ID-modality are ignored. 3) Existing studies do not consider the impact of irrelevant semantic features, leading to inaccurate semantic representation. To address these challenges, we introduce federated semantic learning and devise FFMSR as our solution. For Limitation 1, we locally learn items'semantic encodings from their original texts by a multi-layer semantic encoder, and then cluster them on the server to facilitate the transfer of semantic knowledge between domains. To tackle Limitation 2, we integrate both ID and Text modalities on the clients, and utilize them to learn different aspects of items. To handle Limitation 3, a Fast Fourier Transform (FFT)-based filter and a gating mechanism are developed to alleviate the impact of irrelevant semantic information in the local model. We conduct extensive experiments on two real-world datasets, and the results demonstrate the superiority of our FFMSR method over other SOTA methods. Our source codes are publicly available at: https://github.com/Sapphire-star/FFMSR.

Problem

• Enhancing privacy-preserving cross-domain recommendation under non-overlapping constraints
• Improving semantic representation by exploiting original text and ID modalities
• Reducing irrelevant semantic feature impact with FFT-based filtering

Innovation

• Federated semantic learning for cross-domain recommendations
• Multi-layer semantic encoder for original text encoding
• FFT-based filter and gating for irrelevant feature removal