#data-fusion News & Analysis

2 articles tagged with #data-fusion. AI-curated summaries with sentiment analysis and key takeaways from 50+ sources.

2 articles

AINeutralarXiv – CS AI · Jun 25/10

🧠

Cooperation of Experts: Fusing Heterogeneous Information with Large Margin

Researchers propose the Cooperation of Experts (CoE) framework for fusing heterogeneous data types across different semantic spaces using multiplex networks. The approach employs domain-specific expert encoders that collaborate through a large margin mechanism, demonstrating superior performance across diverse benchmarks with theoretical guarantees on stability and feasibility.

AINeutralarXiv – CS AI · Feb 274/105

🧠

Modality-Guided Mixture of Graph Experts with Entropy-Triggered Routing for Multimodal Recommendation

Researchers introduce MAGNET, a new AI system for multimodal recommendation that combines user behavior, visual, and textual data through specialized graph neural network experts. The system uses entropy-triggered routing to automatically balance different data types and improve recommendations for sparse datasets and long-tail items.