Enabling Incremental Knowledge Transfer for Object Detection at the Edge

Cross-layer host-network co-design for QoS in streamed video data networking

“…Farhadi et al [42] introduce an object detection framework that employs incremental knowledge transfer at the edge. This represents a pioneering effort towards a structured approach for server-assisted online learning.…”

“…Continuous learning aims at tackling the problem of concept drift. Continuous learning approaches have been adopted in existing video analytics designs to refine DNN models continuously online [113,42,152,94,38]. Most existing designs however mainly focus on training techniques without considering systematic design issues such as multi-stream resource contention.…”