Ph.D. student Chao Ma's paper "Towards Making Unsupervised Graph Hashing Discriminative" (by Chao Ma, Chen Gong, Xiang Li, Xiaolin Huang ，Jie Yang) was accepted by IEEE Transactions on Multimedia (TMM).

      The paper is innovative in the following aspects:            

      1. In the process of hashing, not only the similarity between samples is considered, but also the mapping of dissimilar samples to dissimilar hash codes is considered.            

      2. In order to improve the efficiency of measuring the similarity of samples in the original space, the adjacency matrix is constructed by random projection tree and neighbor diffusion.            

      3. In order to solve the problem of sample congestion in hash space, a long tail distribution is proposed to measure the distance between samples, and to maximize the probability that similar samples are projected into similar hash codes and that similar samples are projected into dissimilar hash codes.