To establish a sound traceability system of agricultural products and guarantee security of agricultural products,an algorithm is proposed to encrypt trace code of agricultural products.Original trace code consists of...To establish a sound traceability system of agricultural products and guarantee security of agricultural products,an algorithm is proposed to encrypt trace code of agricultural products.Original trace code consists of 34 digits indicating such information as place of origin,name of product,date of production and authentication.Area code is used to indicate enterprise information,the encrypted algorithm is designed because of the increasing code length,such coding algorithms as system conversion and section division are applied for the encrypted conversion of code of origin place and production date code,moreover,section identification code and authentication code are permutated and combined to produce check code.Through the multiple encryption and code length compression,34 digits are compressed to 20 on the basis of ensuring complete coding information,shorter code length and better encryption enable the public to know information about agricultural products without consulting professional database.展开更多
针对以往医疗对话生成方法未能有效建模医学知识,导致生成的回复缺乏医学常识一贯性的问题,很多学者尝试引入医疗知识图谱,但集成医疗知识图谱时容易占用较多输入数据空间,这限制了模型输入可以保留对话上下文信息量的大小。本文提出知...针对以往医疗对话生成方法未能有效建模医学知识,导致生成的回复缺乏医学常识一贯性的问题,很多学者尝试引入医疗知识图谱,但集成医疗知识图谱时容易占用较多输入数据空间,这限制了模型输入可以保留对话上下文信息量的大小。本文提出知识嵌入的医疗对话生成模型(medical conversation generation model based on knowledge embedding,MCG-KE),该模型基于历史对话进行实体预测得到上下文知识嵌入实体,引入串行图编码方式和图注意力机制获得当前对话相关的医疗知识图谱子图编码,将上下文知识嵌入实体、医疗知识图谱子图编码和历史对话编码作为对话生成模型的输入,用于知识嵌入的医疗对话生成。实验结果表明,模型在高效计算的情况下,所生成的医疗对话在自动评价和人工评价等相关指标上的性能均有提升。展开更多
基金Supported by National 863 Program(2008AA10Z220)Natural Science Research Program of Education Department of Henan Province(2007520020)
文摘To establish a sound traceability system of agricultural products and guarantee security of agricultural products,an algorithm is proposed to encrypt trace code of agricultural products.Original trace code consists of 34 digits indicating such information as place of origin,name of product,date of production and authentication.Area code is used to indicate enterprise information,the encrypted algorithm is designed because of the increasing code length,such coding algorithms as system conversion and section division are applied for the encrypted conversion of code of origin place and production date code,moreover,section identification code and authentication code are permutated and combined to produce check code.Through the multiple encryption and code length compression,34 digits are compressed to 20 on the basis of ensuring complete coding information,shorter code length and better encryption enable the public to know information about agricultural products without consulting professional database.
文摘针对以往医疗对话生成方法未能有效建模医学知识,导致生成的回复缺乏医学常识一贯性的问题,很多学者尝试引入医疗知识图谱,但集成医疗知识图谱时容易占用较多输入数据空间,这限制了模型输入可以保留对话上下文信息量的大小。本文提出知识嵌入的医疗对话生成模型(medical conversation generation model based on knowledge embedding,MCG-KE),该模型基于历史对话进行实体预测得到上下文知识嵌入实体,引入串行图编码方式和图注意力机制获得当前对话相关的医疗知识图谱子图编码,将上下文知识嵌入实体、医疗知识图谱子图编码和历史对话编码作为对话生成模型的输入,用于知识嵌入的医疗对话生成。实验结果表明,模型在高效计算的情况下,所生成的医疗对话在自动评价和人工评价等相关指标上的性能均有提升。