Neural Networks Power aware inference convolutional neural networks semantic segmentation
As Neural Networks are deployed in edge devices, design choices are often dictated by power and compute limitations despite variability in these factors. In this work we propose a novel method for adaptive neural network architectures to boost performance as power becomes available without necessitating a complete reconfiguration of model parameters. We show that due to the stochastic nature of neural networks, models can be constructed with enough independence to be effectively ensembled while sharing a common convolutional base. This allows for a trade-off between power consumption and model accuracy without redeploying an entirely new model. Our method is agnostic to the base network and can be added to existing networks with minimal retraining. We find this approach particularly well suited for IoT devices and distributed autonomous applications where available power and compute resources are time varying.