An energy-efficient memory-based high-throughput VLSI architecture for convolutional networks | IEEE Conference Publication | IEEE Xplore