Towards Understanding Model Quantization for Reliable Deep Neural Network Deployment | IEEE Conference Publication | IEEE Xplore