Accelerating Matrix Multiplication in Deep Learning by Using Low-Rank Approximation | IEEE Conference Publication | IEEE Xplore