“Tongguan Liqiao”acupuncture method established by Xuemin Shi,a master of traditional Chinese medicine in China and academician of the Chinese Academy of Engineering,derives from the famous“Xing Nao Kai Qiao”thera...“Tongguan Liqiao”acupuncture method established by Xuemin Shi,a master of traditional Chinese medicine in China and academician of the Chinese Academy of Engineering,derives from the famous“Xing Nao Kai Qiao”therapy(activating the brain and opening the orifices)acupuncture method.It fundamentally takes“adjusting the spirit”as the root of selecting acupoints based on syndrome differentiation.Neiguan(PC 6),Renzhong(DU 26),and Sanyinjiao(SP 6)are selected as the main acupoints to adjust the spirit and revive the brain,nourish the liver and kidney,which treats the root cause.Acupoints such as Lianquan(RN 23),Yifeng(SJ 17),Wangu(GB 12),and Fengchi(GB 20)in the neck region are selected for“unblocking orifices,”which relieves the symptoms.This acupuncture method features deep needling,which aligns with modern swallowing function anatomy.The method can improve the function of poststroke dysphagia by enhancing cerebral blood supply and metabolism and promoting neural functional remodeling.展开更多
针对关系抽取(RE)任务中实体关系语义挖掘困难和预测关系有偏差等问题,提出一种基于掩码提示与门控记忆网络校准(MGMNC)的RE方法。首先,利用提示中的掩码学习实体之间在预训练语言模型(PLM)语义空间中的潜在语义,通过构造掩码注意力权...针对关系抽取(RE)任务中实体关系语义挖掘困难和预测关系有偏差等问题,提出一种基于掩码提示与门控记忆网络校准(MGMNC)的RE方法。首先,利用提示中的掩码学习实体之间在预训练语言模型(PLM)语义空间中的潜在语义,通过构造掩码注意力权重矩阵,将离散的掩码语义空间相互关联;其次,采用门控校准网络将含有实体和关系语义的掩码表示融入句子的全局语义;再次,将它们作为关系提示校准关系信息,随后将句子表示的最终表示映射至相应的关系类别;最后,通过更好地利用提示中掩码,并结合传统微调方法的学习句子全局语义的优势,充分激发PLM的潜力。实验结果表明,所提方法在SemEval(SemEval-2010 Task 8)数据集的F1值达到91.4%,相较于RELA(Relation Extraction with Label Augmentation)生成式方法提高了1.0个百分点;在SciERC(Entities, Relations, and Coreference for Scientific knowledge graph construction)和CLTC(Chinese Literature Text Corpus)数据集上的F1值分别达到91.0%和82.8%。所提方法在上述3个数据集上均明显优于对比方法,验证了所提方法的有效性。相较于基于生成式的方法,所提方法实现了更优的抽取性能。展开更多
基金the National Key Research and Development Program of China(2018YFC1705004)Tianjin Science and Technology Plan Project(21JCZDJC00890)National Natural Science Foundation of China(82374578).
文摘“Tongguan Liqiao”acupuncture method established by Xuemin Shi,a master of traditional Chinese medicine in China and academician of the Chinese Academy of Engineering,derives from the famous“Xing Nao Kai Qiao”therapy(activating the brain and opening the orifices)acupuncture method.It fundamentally takes“adjusting the spirit”as the root of selecting acupoints based on syndrome differentiation.Neiguan(PC 6),Renzhong(DU 26),and Sanyinjiao(SP 6)are selected as the main acupoints to adjust the spirit and revive the brain,nourish the liver and kidney,which treats the root cause.Acupoints such as Lianquan(RN 23),Yifeng(SJ 17),Wangu(GB 12),and Fengchi(GB 20)in the neck region are selected for“unblocking orifices,”which relieves the symptoms.This acupuncture method features deep needling,which aligns with modern swallowing function anatomy.The method can improve the function of poststroke dysphagia by enhancing cerebral blood supply and metabolism and promoting neural functional remodeling.
文摘针对关系抽取(RE)任务中实体关系语义挖掘困难和预测关系有偏差等问题,提出一种基于掩码提示与门控记忆网络校准(MGMNC)的RE方法。首先,利用提示中的掩码学习实体之间在预训练语言模型(PLM)语义空间中的潜在语义,通过构造掩码注意力权重矩阵,将离散的掩码语义空间相互关联;其次,采用门控校准网络将含有实体和关系语义的掩码表示融入句子的全局语义;再次,将它们作为关系提示校准关系信息,随后将句子表示的最终表示映射至相应的关系类别;最后,通过更好地利用提示中掩码,并结合传统微调方法的学习句子全局语义的优势,充分激发PLM的潜力。实验结果表明,所提方法在SemEval(SemEval-2010 Task 8)数据集的F1值达到91.4%,相较于RELA(Relation Extraction with Label Augmentation)生成式方法提高了1.0个百分点;在SciERC(Entities, Relations, and Coreference for Scientific knowledge graph construction)和CLTC(Chinese Literature Text Corpus)数据集上的F1值分别达到91.0%和82.8%。所提方法在上述3个数据集上均明显优于对比方法,验证了所提方法的有效性。相较于基于生成式的方法,所提方法实现了更优的抽取性能。
