No abstract

No abstract

Recently, deep generative models have become increasingly popular in unsupervised anomaly detection. However, deep generative models aim at recovering the data distribution rather than detecting anomalies. Moreover, deep generative models have the risk of overfitting training samples, which has disastrous effects on anomaly detection performance. To solve the above two problems, we propose a self-adversarial variational autoencoder (adVAE) with a Gaussian anomaly prior assumption. We assume that both the anomalous and the normal prior distribution are Gaussian and have overlaps in the latent space. Therefore, a Gaussian transformer net T is trained to synthesize anomalous but near-normal latent variables. Keeping the original training objective of a variational autoencoder, a generator G tries to distinguish between the normal latent variables encoded by E and the anomalous latent variables synthesized by T, and the encoder E is trained to discriminate whether the output of G is real. These new objectives we added not only give both G and E the ability to discriminate, but also become an additional regularization mechanism to prevent overfitting. Compared with other competitive methods, the proposed model achieves significant improvements in extensive experiments. The employed datasets and our model are available in a Github repository.

No abstract

How to cite:Please refer to published version for the most recent bibliographic citation information.

No abstract

No abstract

With the global commercialization of the fifth-generation (5G) network, many countries, including China, USA, European countries, Japan, and Korea, have started exploring 6G mobile communication network, following the tradition of "planning the next while commercializing one generation". Currently, studies on 6G networks are at the infancy stage. Research on the vision and requirements for 6G is still ongoing, and the industry is yet to clarify the key enabling technologies for 6G. However, 6G will certainly build on the success of 5G. Therefore, developing high-quality 5G networks and seamlessly integrating 5G with verticals are the priorities before 2030, when 6G is projected to be commercialized. Also, global 5G standards will keep evolving to better support vertical applications. As a milestone, the Third-Generation Partnership Project (3GPP) published Release 16 in July 2020, which continuously enhanced the capabilities of mobile broadband service based on Release 15 and realized the support for low-delay and high-reliability applications, such as Internet of Vehicles and industrial Internet. Currently, 3GPP is working on Releases 17 and 18, focusing on meeting the demands of medium-and high-data-rate machine communication with low-cost and high-precision positioning, which will be published in June 2022. Thus, 6G networks will further expand the application fields and scope of the Internet of Things to accommodate those services and applications that are beyond the capabilities of 5G networks. Herein, we present our vision, application scenarios, and key technological trends for 6G networks. Furthermore, we propose several future research opportunities in 6G networks with regard to industrialization and standardization.