Accelerating Deep Neural Network Tasks Through Edge-Device Adaptive Inference | IEEE Conference Publication | IEEE Xplore