LRADNN: High-throughput and energy-efficient Deep Neural Network accelerator using Low Rank Approximation | IEEE Conference Publication | IEEE Xplore