no code implementations • 4 Jan 2019 • Rongrong Ma, Jianyu Miao, Lingfeng Niu, Peng Zhang
To this end, we introduce a new non-convex integrated transformed $\ell_1$ regularizer to promote sparsity for DNNs, which removes both redundant connections and unnecessary neurons simultaneously.
no code implementations • 19 Nov 2018 • Xiao-Bo Jin, Kai-Zhu Huang, Jianyu Miao
AEEN-HOA consists of two branches, i. e., the upper stream is capable of erasing some initially discovered regions, then the high-order attribute supervision is incorporated to characterize the relationship between the class attributes.