Reinforcement Learning-Based Layer-Wise Quantization For Lightweight Deep Neural Networks | IEEE Conference Publication | IEEE Xplore