Recently, the gold suit ant algorithm engineer Hu Binbin wrote essays "Adversarial Learning on Heterogeneous Information Networks" was selected Global Data Mining detailed interpretation of the premier meeting KDD 2019, for the thesis of this article. Papers Address:
https://www.kdd.org/kdd2019/accepted-papers/view/adversarial-learning-on-heterogeneous-information-networks
Foreword
Learning network representation is a representation of network data in a low dimensional space method has been widely used in heterogeneous information network analysis. Existing heterogeneous information network representation learning method, although to achieve improved performance to a certain extent, but there are still some major deficiencies. Most importantly, they usually employed method negative samples randomly selected node from the network, rather than the underlying distribution study to obtain a more robust representation.
By type of generation