Research
The group currently has four research directions: Network Verification, Programmable Networks, Transmission Performance Optimization, and Explainable AI. Click the icons for detailed introductions.
NetVerify
Network Verification
As network scale and complexity increase, especially with the emergence of large-scale training clusters, traditional network operations relying on manual experience are no longer suitable. Network verification technology formalizes and accurately simulates network topology, configuration, and traffic states, enabling automated generation, verification, and repair of networks to ensure reliability. Network verification has become an indispensable key technology for major cloud providers, operators, and internet companies, with significant research and application value.
NetFun
Programmable Networks
Programmable networks leverage technologies such as Software-Defined Networking (SDN) for flexible programming control over network behavior, dynamically adjusting network resources and policies to effectively enhance network agility and accelerate the evolution of network architecture.
NetPerf
Transmission Performance Optimization
Network transmission performance optimization aims to reduce latency, improve bandwidth utilization, and enhance data transmission efficiency. By employing techniques such as traffic management, transmission control, and intelligent scheduling, it minimizes congestion and data loss, thereby boosting network stability and responsiveness. Research in this area contributes to the development of high-performance networks that meet the demands for high-speed communication in modern applications.
XAI
Explainable AI
The core goal of Explainable AI is to enhance the transparency and trustworthiness of AI systems. Through innovative techniques such as visual analytics, feature attribution, and decision path tracing, it deeply analyzes the reasoning mechanisms of models, effectively addressing the traditional "black box" problem of AI. This research is of key significance in high-value application scenarios such as medical diagnosis, financial risk control, and autonomous driving, ensuring that AI decision-making processes meet fairness, compliance, and traceability requirements.