<ul><li>Data synthesis is gaining momentum as a privacy-enhancing technology, with a focus on multi-table data generation that poses challenges in capturing complex relational structures.</li><li>Current methods for multi-table data struggle with long-range dependencies and complex foreign-key relationships, such as tables with multiple parent tables or various types of links between the same pair of tables.</li><li>A new generative model for relational data is proposed, generating relational datasets based on the graph formed by foreign-key relationships through flow matching.</li><li>The method uses a deep generative model and a graph neural network to denoise records and extract information from connected records, achieving state-of-the-art performance in generating synthetic data.</li></ul>

Graph Conditional Flow Matching for Relational Data Generation

Discover more