Search

Distributed denial of service (DDoS) attacks pose a huge threat to the Internet. Follow the rapid usage of the Internet of things (IoT), the DDoS attack is no longer a mere traffic attack, the original attack on the application layer surpasses the attack on the network layer. Furthermore, DDoS attacks using Bonnets result more destructive effects. This research aims to propose a new collaborated active defense framework between Honeypot and cloud platform to detect and defend future DDoS attacks in the context of the IoT with the instantaneous malicious traffic measured in Terabytes.

In recent years, with the development of network technology and the development of real estate industry, intelligent apartments are becoming more and more popular. The intelligent door lock equipment is becoming more and more popular. Compared with the traditional door lock, the intelligent door lock has higher security and convenience. The user controls the door lock through the mobile device, and the apartment controls the door lock through the computer. Compared with the traditional apartment door lock or electronic access card, it is more convenient. Therefore, a management system that can manage the door locks intelligently through the Internet is of great significance to the real estate industry and social development. In this paper, using the existing knowledge of electronic door lock in the industry, combined with the author's professional ability, an intelligent door lock management system based on raspberry pie, Java EE and Android is designed. The door lock end is controlled by raspberry pie, which scans fingerprint data and enters fingerprint data by as608 fingerprint module. Rpi-lcd35b display screen is used as password input interface, RPI camera is used as monitoring equipment of door lock. Raspberry pie controls electronic door lock and sends door opening record to server. Users can get the information of the door lock through the mobile app, modify the door lock password, save the household information and perform other operations, which can be monitored by the camera on the door lock.

Perceiving user emotions and generating responses with specific emotions are of great significance to a social dialogue system. Much previous work still utilizes seq2seq models as the backbone to generate emotional responses. With the popularity of the generative pre-training models, we propose an emotional response generation model EmoDialoGPT based on DialoGPT by introducing emotion embedding and emotion prediction loss. In order to obtain a large-scale dialogue dataset with emotion labels, we train an emotion classifier and automatically annotate the emotion labels for the dialogue data. We evaluate our models from three aspects, including emotion expression, response quality and human evaluation. The experimental results show that, the proposed models outperform seq2seq models and baseline DialoGPT models on most metrics.

Drug information includes indications, contraindications, adverse reactions and other specific descriptions, which can provide a basis and support for rational clinical rational drug use and clinical decision-making. A Chinese medicine knowledge base with interrelated disease and drug information interrelated can effectively monitor the therapeutic effects of drugs on diseases and provide knowledge resources for drug research. In this paper, the multi-source Chinese Medicine Knowledge Base (CMKB) is constructed in view of the important role of drug information in clinical medication, auxiliary diagnosis, disease risk prediction and other practical applications. CMKB takes the mainstream medical resources such as the China Disease Knowledge Total Database and Dingxiangyuan as data sources. On the basis of the knowledge description system and classification system of the medicine database, CMKB standardizes the classification of drugs from multiple sources and forms corresponding detailed knowledge descriptions. CMKB constructed contains 14,141 drugs, which are described in detail from 7 aspects such as indications, contraindications and classification. Knowledge association between drug entities and disease entities is established through named entity recognition. The accuracy of named entity recognition reaches 90.6% and 91.4% as assessed by different models. CMKB constructed in this paper can be a useful supplement for the Chinese Medical Knowledge Graph (CMeKG) and will provide support for intelligent diagnosis and clinical decision-making.