In recent years, artificial intelligence technologies have been widely used in computer vision, natural language processing, automatic driving, and other fields. However, artificial intelligence systems are vulnerable to adversarial attacks, which limit the applications of artificial intelligence (AI) technologies in key security fields. Therefore, improving the robustness of AI systems against adversarial attacks has played an increasingly important role in the further development of AI. This paper aims to comprehensively summarize the latest research progress on adversarial attack and defense technologies in deep learning. According to the target model’s different stages where the adversarial attack occurred, this paper expounds the adversarial attack methods in the training stage and testing stage respectively. Then, we sort out the applications of adversarial attack technologies in computer vision, natural language processing, cyberspace security, and the physical world. Finally, we describe the existing adversarial defense methods respectively in three main categories, i.e., modifying data, modifying models and using auxiliary tools.

No abstract

Heavy smoking can induce airway inflammation and emphysema. Macrolides can modulate inflammation and effector T-cell response in the lungs. However, there is no information on whether erythromycin can modulate regulatory T-cell (Treg) response. This study is aimed at examining the impact of erythromycin on Treg response in the lungs in a rat model of smoking-induced emphysema. Male Wistar rats were exposed to normal air or cigarette smoking daily for 12 weeks and treated by gavage with 100 mg/kg of erythromycin or saline daily beginning at the forth week for nine weeks. The lung inflammation and the numbers of inflammatory infiltrates in bronchoalveolar lavage fluid (BALF) were characterized. The frequency, the number of Tregs, and the levels of Foxp3 expression in the lungs and IL-8, IL-35, and TNF-α in BALF were determined by flow cytometry, RT-PCR and ELISA, respectively. Treatment with erythromycin reduced smoking-induced inflammatory infiltrates, the levels of IL-8 and TNF-α in the BALF and lung damages but increased the numbers of CD4+Foxp3+ Tregs and the levels of Foxp3 transcription in the lungs, accompanied by increased levels of IL-35 in the BALF of rats. Our novel data indicated that erythromycin enhanced Treg responses, associated with the inhibition of smoking-induced inflammation in the lungs of rats.