Cloud Learning-by-Education Node Community (C-LENC) framework

The Learning-by-Education Node Community (LENC) paradigm has laid out the foundation for examining the interaction rules between collaborative nodes performing knowledge exchange, based on evaluating their knowledge. As a research prototype, Out-Of-Distribution detectors were employed for knowledge (self-) assessment and knowledge distillation for model updates, all operating on the same computer. This paper presents an extension to the LENC paradigm, the novel Cloud Learning-by-Education Node Community (C-LENC) framework, that can be used to perform a plethora of collaborative distributed machine learning workflows between the nodes of the C-LENC network, operating on the cloud. It leverages a succinct communication protocol, exposing a small number of commands the C-LENC nodes can utilize to exchange information. We demonstrate the usage of C-LENC on four distributed machine learning workflows, including learning a new task, federated learning, multi-teacher knowledge distillation and distributed inference. We also present the internal architecture of a LENC node, as well as the procedure for utilizing Out-Of-Distribution detectors to query the C-LENC network and finding suitable CLENC nodes for any task. We present the limitations as well as potential directions to further improve the C-LENC framework.