Basis of seventy-seven case records of twenty-six diseases in Linzheng Zhinan Yi'an(Guide to Clinical Practice),analyze and distinguish them into four types.They are disorder of the mother-organ affecting the chil...Basis of seventy-seven case records of twenty-six diseases in Linzheng Zhinan Yi'an(Guide to Clinical Practice),analyze and distinguish them into four types.They are disorder of the mother-organ affecting the child-organ,disorder of the child-organ affecting the mother-organ,over restraint and reverse restraint.The essence of this article is to recognize the pathological transmission among the Zang-fu organs according to the theory of five elements.展开更多
在法律判决预测领域中,案件描述通常具有相似的结构,而现有的预测方法容易忽略不同案件间的要素差异,难以有效利用案件要素特征,导致模型预测准确率不高.此外,罪名预测任务还面临易混淆罪名问题.针对上述问题,提出了一种融合案件要素和...在法律判决预测领域中,案件描述通常具有相似的结构,而现有的预测方法容易忽略不同案件间的要素差异,难以有效利用案件要素特征,导致模型预测准确率不高.此外,罪名预测任务还面临易混淆罪名问题.针对上述问题,提出了一种融合案件要素和案件属性的罪名预测多任务学习模型(Case Elements And Attributes Multi-Task Learning Model,简称CEAT-MLM),通过挖掘案件要素及案件属性与罪名之间的关联关系,将案件属性预测和罪名预测进行联合建模,达到提升罪名预测准确率的目标.实验结果表明,本文提出的模型相较于通用的文本分类模型具有显著的性能提升,并与法律判决领域的典型模型相比,Macro-F1得分提升了1.76%.展开更多
The nature of farmer cooperative economy organization( known as FCEO) determines the fact that the economic effects of farmer cooperative economy organization are as important as its social effects. Many experts,howev...The nature of farmer cooperative economy organization( known as FCEO) determines the fact that the economic effects of farmer cooperative economy organization are as important as its social effects. Many experts,however,now would only focus on its economic function, and either neglect or weaken its social influence. Therefore,this paper introduces the theoretical foundation of the farmer cooperative economy organization,and studies the nature of cooperative economics. Based on those typical cases,the future of cooperative organization and four supporting elements were put forward in this paper.展开更多
针对现有的类案检索(LCR)方法缺乏对案情要素的有效利用而容易被案例内容的语义结构相似性误导的问题,提出一种融合时序行为链与事件类型的类案检索方法。首先,采取序列标注的方法识别案情描述中的法律事件类型,并利用案例文本中的行为...针对现有的类案检索(LCR)方法缺乏对案情要素的有效利用而容易被案例内容的语义结构相似性误导的问题,提出一种融合时序行为链与事件类型的类案检索方法。首先,采取序列标注的方法识别案情描述中的法律事件类型,并利用案例文本中的行为要素构建时序行为链,以突出案情的关键要素,从而使模型聚焦于案例的核心内容,进而解决现有方法易被案例内容的语义结构相似性误导的问题;其次,利用分段编码构造时序行为链的相似性向量表征矩阵,从而增强案例间行为要素的语义交互;最后,通过聚合评分器,从时序行为链、法律事件类型、犯罪类型这3个角度衡量案例的相关性,从而增加案例匹配得分的合理性。实验结果表明,相较于SAILER(Structure-Aware pre-traIned language model for LEgal case Retrieval)方法,所提方法在LeCaRD(Legal Case Retrieval Dataset)上的P@5值提升了4个百分点、P@10值提升了3个百分点、MAP值提升了4个百分点,而NDCG@30值提升了0.8个百分点。可见,该方法能有效利用案情要素来避免案例内容的语义结构相似性的干扰,并能为类案检索提供可靠的依据。展开更多
文摘Basis of seventy-seven case records of twenty-six diseases in Linzheng Zhinan Yi'an(Guide to Clinical Practice),analyze and distinguish them into four types.They are disorder of the mother-organ affecting the child-organ,disorder of the child-organ affecting the mother-organ,over restraint and reverse restraint.The essence of this article is to recognize the pathological transmission among the Zang-fu organs according to the theory of five elements.
文摘在法律判决预测领域中,案件描述通常具有相似的结构,而现有的预测方法容易忽略不同案件间的要素差异,难以有效利用案件要素特征,导致模型预测准确率不高.此外,罪名预测任务还面临易混淆罪名问题.针对上述问题,提出了一种融合案件要素和案件属性的罪名预测多任务学习模型(Case Elements And Attributes Multi-Task Learning Model,简称CEAT-MLM),通过挖掘案件要素及案件属性与罪名之间的关联关系,将案件属性预测和罪名预测进行联合建模,达到提升罪名预测准确率的目标.实验结果表明,本文提出的模型相较于通用的文本分类模型具有显著的性能提升,并与法律判决领域的典型模型相比,Macro-F1得分提升了1.76%.
基金Supported by the Youth Program of Chongqing Social Science Plan(No.2012QNGL047)West Program of Humanistic and Social Science of Education Department(No.13XJC630006)+1 种基金Education and Teaching Program of Southwest University(No.2012JY037)Chongqing Science Committee Decision-making Subject(No.2013KXKT07)
文摘The nature of farmer cooperative economy organization( known as FCEO) determines the fact that the economic effects of farmer cooperative economy organization are as important as its social effects. Many experts,however,now would only focus on its economic function, and either neglect or weaken its social influence. Therefore,this paper introduces the theoretical foundation of the farmer cooperative economy organization,and studies the nature of cooperative economics. Based on those typical cases,the future of cooperative organization and four supporting elements were put forward in this paper.
文摘针对现有的类案检索(LCR)方法缺乏对案情要素的有效利用而容易被案例内容的语义结构相似性误导的问题,提出一种融合时序行为链与事件类型的类案检索方法。首先,采取序列标注的方法识别案情描述中的法律事件类型,并利用案例文本中的行为要素构建时序行为链,以突出案情的关键要素,从而使模型聚焦于案例的核心内容,进而解决现有方法易被案例内容的语义结构相似性误导的问题;其次,利用分段编码构造时序行为链的相似性向量表征矩阵,从而增强案例间行为要素的语义交互;最后,通过聚合评分器,从时序行为链、法律事件类型、犯罪类型这3个角度衡量案例的相关性,从而增加案例匹配得分的合理性。实验结果表明,相较于SAILER(Structure-Aware pre-traIned language model for LEgal case Retrieval)方法,所提方法在LeCaRD(Legal Case Retrieval Dataset)上的P@5值提升了4个百分点、P@10值提升了3个百分点、MAP值提升了4个百分点,而NDCG@30值提升了0.8个百分点。可见,该方法能有效利用案情要素来避免案例内容的语义结构相似性的干扰,并能为类案检索提供可靠的依据。
