On May 30, 2025, Dr. Yu Peng, Vice Dean of the Future School at Beijing University of Posts and Telecommunications (BUPT), was invited by the School of Artificial Intelligence and Software Engineering to present an academic report titled “6G Edge Network Fault Tolerance: Background, Challenges and Future Directions”. The lecture was chaired by Lan Yihua, Vice Dean of the School, with faculty members, undergraduate and graduate student representatives in attendance.

In his presentation, Dr. Yu provided an in-depth analysis of the challenges facing 6G edge networks and highlighted two innovative solutions developed by his team: 1) A deep reinforcement learning-based fault-tolerant routing mechanism that integrates Message Passing Neural Networks (MPNN) with Deep Q-Networks (DQN) to enable intelligent routing decisions in dynamic environments; 2) A preemptive migration fault-tolerance model for resource management that employs Graph Attention Networks (GAT) and Transformer architectures for precise fault prediction, combined with diffusion models to generate optimal migration strategies. These solutions offer intelligent approaches to enhancing 6G edge network reliability from both routing optimization and resource management perspectives. During the Q&A session, Dr. Yu engaged in in-depth exchanges with attendees, generously sharing his research insights and valuable experience in project proposal development, academic writing, and the manuscript submission process, providing detailed responses to audience questions.

This event has not only strengthened academic collaboration between our university and BUPT in artificial intelligence but also injected new momentum into interdisciplinary integration and innovation, providing substantial support for the development of emerging engineering disciplines and the cultivation of high-quality interdisciplinary talent.

Written & Photo by: Xu Jiashu & Guo Kangda

Source: NYNU Academic Activities (Chinese)

https://www.nynu.edu.cn/info/1048/29927.htm