Double-Win Quant: Aggressively Winning Robustness of Quantized Deep
    Neural Networks via Random Precision Training and Inference

            Yonggan Fu 1 Qixuan Yu 1 Meng Li 2 Vikas Chandra 2 Yingyan Lin 1

            Abstract                Wu et al., 2018). However, DNNs’ deployments into IoT
   Quantization is promising in enabling powerful       devices still remain challenging. First, powerful DNNs often
   yet complex deep neural networks (DNNs) to         come at a prohibitive cost,  ...