A 44.1TOPS/W Precision-Scalable Accelerator for Quantized Neural Networks in 28nm CMOS | IEEE Conference Publication | IEEE Xplore