Network Systems for Distributed Learning
We design network architectures that optimize the efficiency of distributed learning training and inference across heterogeneous and large-scale systems.
We design network architectures that optimize the efficiency of distributed learning training and inference across heterogeneous and large-scale systems.
We leverage software-defined networking to develop flexible telemetry and measurement frameworks that enhance network visibility, adaptability, and overall performance.
We envision space-based datacenter networks that extend beyond connectivity, providing intelligent computing, resource coordination, and integrated data processing capabilities.
We design communication technologies and protocols that efficiently utilize limited spectrum resources to jointly support reliable communication and sensing services.