معرفی
Dr. Dongjin Song is an Assistant Professor in the School of Computing at the University of Connecticut, specializing in continual learning, graph representation, and time series analysis. He holds a PhD from UC San Diego (2016) and previously worked at NEC Labs America. His NSF CAREER award (2024) supports research on evolving graph learning for applications in healthcare, energy, and transportation. Recognized with the Frontiers of Science Award (2024) and UConn AAUP Excellence Award (2025), he develops algorithms addressing catastrophic forgetting and generalization in dynamic systems.
Research interests include:
- Robust representation learning for time-series and graph data
- Meta-knowledge distillation for heterogeneous systems
- Privacy-preserving federated learning
- LLM-enhanced multimodal forecasting
Educational initiatives integrate research into AI/ML courses, and outreach targets K-12 STEM engagement. Current projects explore power outage prediction, clinical time-series analysis, and cross-platform mental health monitoring.
