Gia Oriana Santos
Tim Berners-Lee’s Semantic Web vision imagines autonomous software agents leveraging personal data and the Web to collaborate on complex tasks. Despite significant progress, we remain far from achieving this vision. Knowledge graphs (KGs) and multi-agent systems (MAS) have enabled structured knowledge, transparency, and agent collaboration, but challenges in data privacy, interoperability, and explainability persist. Federated learning (FL) offers promising mechanisms for privacy-preserving knowledge sharing across distributed personal and domain-specific KGs. This survey reviews the state of research at the intersection of KGs, FL, and MAS, highlighting how their integration can address key barriers to privacy, semantic interoperability, and agent reasoning. We identify open challenges and propose future research directions toward realizing the Semantic Web vision.
Links:
- Final paper: https://drive.google.com/file/d/1CZrL2bvzZzteowKVFy4sV3vSEUw1CWhL/
- Final presentation (slides): https://docs.google.com/presentation/d/123Ct6cDqjFWDY_F2_OolhD_hr0fLIY-mVxgoc0rhs9U/
- Final presentation (video): https://youtu.be/7dusO5cFehk
