Multi-Granularity Subgraph Guided Replay for Robust Graph Continual Learning via Granular-Ball Computing
Level: IPM Journal under review
My current research primarily focuses on Continual Learning and Granular-ball Computing. I am dedicated to investigating how machine learning models can continuously absorb new knowledge from dynamic data streams while overcoming the challenge of catastrophic forgetting. Concurrently, I am exploring the potential of granular-ball computing in reducing computational complexity and enhancing model robustness and interpretability, aiming to contribute to the development of more efficient and intelligent AI systems.
Multi-Granularity Subgraph Guided Replay for Robust Graph Continual Learning via Granular-Ball Computing
Funded by the National Natural Science Foundation of China.