![[2602.20877] E-MMKGR: A Unified Multimodal Knowledge Graph Framework for E-commerce Applications](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.20877] E-MMKGR: A Unified Multimodal Knowledge Graph Framework for E-commerce Applications
Summary
The paper presents E-MMKGR, a unified framework for multimodal knowledge graphs tailored for e-commerce, enhancing recommendation systems and product search through improved item representation.
Why It Matters
As e-commerce continues to grow, effective recommendation systems are crucial for user engagement and sales. E-MMKGR addresses limitations in existing multimodal systems, providing a more flexible and effective approach that can adapt to various tasks, thus enhancing user experience and operational efficiency in e-commerce applications.
Key Takeaways
- E-MMKGR improves collaborative filtering by utilizing a multimodal knowledge graph.
- The framework enhances item representation through GNN-based propagation.
- Experiments show significant performance improvements in recommendation and product search.
- The approach allows for greater extensibility and generalization across tasks.
- Real-world applications demonstrate its effectiveness in e-commerce settings.
Computer Science > Information Retrieval arXiv:2602.20877 (cs) [Submitted on 24 Feb 2026] Title:E-MMKGR: A Unified Multimodal Knowledge Graph Framework for E-commerce Applications Authors:Jiwoo Kang, Yeon-Chang Lee View a PDF of the paper titled E-MMKGR: A Unified Multimodal Knowledge Graph Framework for E-commerce Applications, by Jiwoo Kang and Yeon-Chang Lee View PDF HTML (experimental) Abstract:Multimodal recommender systems (MMRSs) enhance collaborative filtering by leveraging item-side modalities, but their reliance on a fixed set of modalities and task-specific objectives limits both modality extensibility and task generalization. We propose E-MMKGR, a framework that constructs an e-commerce-specific Multimodal Knowledge Graph E-MMKG and learns unified item representations through GNN-based propagation and KG-oriented optimization. These representations provide a shared semantic foundation applicable to diverse tasks. Experiments on real-world Amazon datasets show improvements of up to 10.18% in Recall@10 for recommendation and up to 21.72% over vector-based retrieval for product search, demonstrating the effectiveness and extensibility of our approach. Subjects: Information Retrieval (cs.IR); Artificial Intelligence (cs.AI) Cite as: arXiv:2602.20877 [cs.IR] (or arXiv:2602.20877v1 [cs.IR] for this version) https://doi.org/10.48550/arXiv.2602.20877 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Jiwoo Kang [view e...