Multi-modal advertising visual information fusion using dynamic heterogeneous graph neural networks

Key Points

• The aim is to explore how multi-modal advertising can leverage dynamic heterogeneous graph neural networks for better data fusion.
• Developed a dynamic heterogeneous graph neural network framework.
• Integrated various visual information sources from multiple advertising modalities.
• Analyzed the effectiveness of the proposed model with respect to existing methods.
• The proposed model significantly improved advertising effectiveness based on integrated visual data.
• Demonstrated superior performance metrics compared to traditional fusion techniques.
• Achieved higher user engagement rates in multi-modal advertising scenarios.

Abstract