Proteus: Exploiting precision variability in deep neural networks

We evaluate Proteus as an extension to a state-of-the-art accelerator [1] which uses a uniform 16-bit fixed-point representation. On five popular Convolutional Neural Networks (CNNs) and during inference Proteus reduces data traffic by 43% on average while maintaining accuracy within 1% compared to the full precision baseline. As a result, Proteus improves energy by 15% with no performance loss.