Reproduce “Construction and Applications of Billion-Scale Pre-Trained Multimodal Business Knowledge Graph” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Shumin Deng, Chengming Wang, Zhoubo Li, Ningyu Zhang, Zelin Dai, Hehong Chen, Feiyu Xiong, Ming Yan and 6 more
Researchers from Alibaba Group and partner institutions developed OpenBG, a publicly released, billion-scale, multimodal business knowledge graph that standardizes and integrates diverse e-commerce data. This system addresses structural deficiencies and multiple data modalities, demonstrating improved performance in various e-commerce applications, including a 45% increase in Gross Merchandise Volume (GMV) promoted for item alignment and an 11% increase in Click-Through-Rate (CTR) for QA-based product recommendation.
Key insights
- Designing an ontology that is coherent and aligned with specific business logic is fundamental for effectively managing and scaling a large-scale business knowledge graph.
- The integration of knowledge into a business KG should prioritize practicality and minimalism, focusing on information that is directly relevant and valuable to business scenarios.
- Maintaining the quality of a billion-scale knowledge graph is an ongoing process that requires continuous review, refinement, and adaptation over its entire lifecycle.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.