Exploiting Elasticity in Tensor Ranks for Compressing Neural Networks | IEEE Conference Publication | IEEE Xplore