In edge computing, owing to the mobility and power limitation of mobile devices, the edge servers are transformed dynamically. This scenario spawned a new framework called dynamic edge computing. Dynamic edge computing or Mobile Edge Computing (MEC) is the multi-access edge computing that mitigates a large burden from end devices to edge servers.
   Dynamic edge computing requires a more powerful technique to be incorporated due to the challenge of task offloading while transferring tasks in dynamic edge devices. Deep learning models are capable of estimating optimal offloading decisions for the generation task in edge devices.
   Integrating deep learning models in dynamic edge computing assist in offloading computationally intensive tasks for more powerful mobile edge devices to enhance the task offloading reliability. The recent development of deep learning-based task offloading for dynamic edge computing focuses on utilizing distributed deep learning paradigm for a more dynamic environment and different service requirements in edge devices.