At present, people’s demand for the Internet is increasing day by day, and the scale of the Internet is exploding at an incredible speed. This also leads to the increasingly serious problem of routing scalability and mobility. The birth of the LISP network architecture provides an effective method to solve problems stated above. Consistency testing technology is the key technology to ensure the interconnection of network devices under the new network architecture. In this paper, aiming at the typical LISP mapping distribution protocol - LISP-ALT (RFC6836), a conformance testing system is designed and many testcases are implemented by standard TTCN-3 test language and open-source TITAN platform. Finally, a typical testcase is verified and analyzed for demonstration.
Named entity recognition is a key task in the field of natural language processing, which plays a key role in many downstream tasks. Adversarial examples attack based on hard label black box is to generate adversarial examples which make the model classification wrong under the condition that only the decision results of the model are obtained. However, at present, there is little research on adversarial examples attack in hard-label black box setting for named entity recognition task. Influenced by adversarial examples attacks in hard-label black box settings in text classification task, we apply genetic algorithm to adversarial examples attacks in named entity recognition task. In this paper, we first randomly generate the initial adversarial examples, and shorten the search space to a certain extent, and then use genetic algorithm to continuously optimize the examples, and finally generate high quality adversarial examples. Experiments and analysis show that the adversarial examples generated in the hard label black box setting can effectively reduce the accuracy of the model.
The RISC-V instruction set architecture (ISA) stimulates the expeditious development of novel hardware platforms. Consequently, the need for an efficient and easy-to-use operating system on RISC-V architecture emerges. However, new challenges such as system building, hardware adaptation, and application ecosystem should be addressed as the hardware podium develops. This article explores the design and building of openKylin, an open-source operating system, on the RISC-V hardware platform to address these issues, including kernel optimization, UKUI (Ultimate Kylin User Interface) package compilation, and application compatibility. The test results show that the x86 benchmark can run in the openKylin operating system correctly and efficiently on the RISC-V platform.
This paper illustrates the application of an autonomous decentralized system introducing edge nodes in a smart factory. The nodes include atomic nodes, the most fundamental autonomic units constituting an autonomous decentralized system that are capable of independently extracting information from the data field and processing that information internally and simultaneously transmitting the processing results and other internal information to the data field proactively in a broadcast mode, where the transmitted information circulates in the data field. The group-based data field enables the sharing of information by atomic nodes. Grouped management nodes are the managers of the whole group, and edge nodes are the terminal intelligent control units of an intelligent autonomous decentralized system. Improvements in system decentralization, fault tolerance characteristics and flexibility of the system can lead to homogenization, autonomous control and autonomous coordination in complex manufacturing environments, as well as excellent online fault tolerance, online expansion and online maintenance of the system. In addition, it enhances the ever-changing and evolving control requirements by deconstructing the complexity in the system.
The classification and recycling of garbage can greatly improve the utilization of garbage resources. This paper proposes a new convolutional neural network that fuses a multi-branch Xception network with an attention mechanism module. The effective feature information is emphasized and the invalid information is suppressed to overcome the problem caused by the small data set. To verify the usefulness of this network structure in the field of garbage images, this paper uses a widely used data set in the field of garbage image classification. For any network without pre-trained weights, the network proposed in this paper outperforms all other methods by 94.4%.
