Nowadays, deep learning has been increasingly applied in real-world scenarios involving the collection and analysis of sensitive data, which often causes privacy leakage. Differential privacy is widely recognized in the majority of traditional scenarios for its rigorous mathematical guarantee. However, it is uncertain to work effectively in the deep learning model. In this paper, we introduce the privacy attacks facing the deep learning model and present them from three aspects: membership inference, training data extraction, and model extracting. Then we recall some basic theory about differential privacy and its extended concepts in deep learning scenarios. Second, in order to analyze the existing works that combine differential privacy and deep learning, we classify them by the layers differential privacy mechanism deployed, such as input layer, hidden layer, and output layer, and discuss their advantages and disadvantages. Finally, we point out several key issues to be solved and provide a broader outlook of this research direction.INDEX TERMS Deep learning, differential privacy, privacy attacks.
The first rotameric monoterpenoid indole alkaloids (MIAs), 1a and 1b, and two unusual dimeric MIAs, 2 and 3, with new dimerization patterns, together with their putative biosynthetic intermediates 4-7, were isolated from the roots of Gelsemium elegans. Compounds 2 and 3 represent the first natural aromatic azo- and the first urea-linked dimeric MIAs, respectively. Their structures and absolute configurations were elucidated by means of NMR spectroscopy, single-crystal X-ray diffraction, and electronic circular dichroism data analyses. The interconverting mechanism of rotamers 1a and 1b was studied by density functional theory computation. Compounds 2 and 3 showed moderate cytotoxic activity against MCF-7 and PC-12 cells, respectively. In addition, a plausible biosynthesis pathway for the new alkaloids was proposed on the basis of the coexistence of their biosynthetic precursors.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.