Abstract: A paramyxovirus vaccine strain for novel coronavirus pneumonia and a construction method thereof are provided. The method includes performing a recombination of N and F genes of Newcastle disease virus type VII of Paramyxoviridae with P, M, H, L genes of Canine distemper virus of Paramyxoviridae to obtain a recombinant virus, inserting S1 gene of the novel coronavirus between the P and M genes of the recombinant virus to obtain a recombinant vector. The vaccine strain constructed can stably and efficiently express the novel coronavirus S1 protein, and induce the body to produce antibodies; and the recombined virus vaccine strain can stimulate the human body to produce mucosal immunity, and the prepared vaccine can be vaccinated through a nasal spray. Moreover, the vaccine strain can be tested in poultry and dogs, saving time, reducing costs, and being more conducive to actual production due to large output.

Abstract: A paramyxovirus vaccine strain for novel coronavirus pneumonia and a construction method thereof are provided. The method includes performing a recombination of N and F genes of Newcastle disease virus type VII of Paramyxoviridae with P, M, H, L genes of Canine distemper virus of Paramyxoviridae to obtain a recombinant virus, inserting S1 gene of the novel coronavirus between the P and M genes of the recombinant virus to obtain a recombinant vector. The vaccine strain constructed can stably and efficiently express the novel coronavirus S1 protein, and induce the body to produce antibodies; and the recombined virus vaccine strain can stimulate the human body to produce mucosal immunity, and the prepared vaccine can be vaccinated through a nasal spray. Moreover, the vaccine strain can be tested in poultry and dogs, saving time, reducing costs, and being more conducive to actual production due to large output.

Abstract: A long-term memory network method and system for text comprehension. A recurrent neural network can be provided, which includes an external memory module and a long-short term memory unit, wherein said recurrent neural network encodes raw text information into vector representations, forms memories, finds relevant sentences to answer questions, and generates multi-word answers to said questions utilizing the long short term memory unit.

Abstract: A long-term memory network method and system for text comprehension. A recurrent neural network can be provided, which includes an external memory module and a long-short term memory unit, wherein said recurrent neural network encodes raw text information into vector representations, forms memories, finds relevant sentences to answer questions, and generates multi-word answers to said questions utilizing the long short term memory unit.