以编目分类和规则匹配为主的古籍文本主题分类方法存在工作效能低、专家知识依赖性强、分类依据单一化、古籍文本主题自动分类难等问题。对此,本文结合古籍文本内容和文字特征,尝试从古籍内容分类得到符合研究者需求的主题,推动数字人...以编目分类和规则匹配为主的古籍文本主题分类方法存在工作效能低、专家知识依赖性强、分类依据单一化、古籍文本主题自动分类难等问题。对此,本文结合古籍文本内容和文字特征,尝试从古籍内容分类得到符合研究者需求的主题,推动数字人文研究范式的转型。首先,参照东汉古籍《说文解字》对文字的分析方式,以前期标注的古籍语料数据集为基础,构建全新的“字音(说)-原文(文)-结构(解)-字形(字)”四维特征数据集。其次,设计四维特征向量提取模型(speaking,word,pattern,and font to vector,SWPF2vec),并结合预训练模型实现对古籍文本细粒度的特征表示。再其次,构建融合卷积神经网络、循环神经网络和多头注意力机制的古籍文本主题分类模型(dianji-recurrent convolutional neural networks for text classification,DJ-TextRCNN)。最后,融入四维语义特征,实现对古籍文本多维度、深层次、细粒度的语义挖掘。在古籍文本主题分类任务上,DJ-TextRCNN模型在不同维度特征下的主题分类准确率均为最优,在“说文解字”四维特征下达到76.23%的准确率,初步实现了对古籍文本的精准主题分类。展开更多
传统编目分类和规则匹配方法存在工作效能低、过度依赖专家知识、缺乏对古籍文本自身语义的深层次挖掘、编目主题边界模糊、较难实现对古籍文本领域主题的精准推荐等问题。为此,本文结合古籍语料特征探究如何实现精准推荐符合研究者需...传统编目分类和规则匹配方法存在工作效能低、过度依赖专家知识、缺乏对古籍文本自身语义的深层次挖掘、编目主题边界模糊、较难实现对古籍文本领域主题的精准推荐等问题。为此,本文结合古籍语料特征探究如何实现精准推荐符合研究者需求的文本主题内容的方法,以推动数字人文研究的进一步发展。首先,选取本课题组前期标注的古籍语料数据进行主题类别标注和视图分类;其次,构建融合BERT(bidirectional encoder representation from transformers)预训练模型、改进卷积神经网络、循环神经网络和多头注意力机制的语义挖掘模型;最后,融入“主体-关系-客体”多视图的语义增强模型,构建DJ-TextRCNN(DianJi-recurrent convolutional neural networks for text classification)模型实现对典籍文本更细粒度、更深层次、更多维度的语义挖掘。研究结果发现,DJ-TextRCNN模型在不同视图下的古籍主题推荐任务的准确率均为最优。在“主体-关系-客体”视图下,精确率达到88.54%,初步实现了对古籍文本的精准主题推荐,对中华文化深层次、细粒度的语义挖掘具有一定的指导意义。展开更多
Objective To explore whether the protein Deglycase protein 1(DJ1)can ameliorate Alzheimer’s disease(AD)-like pathology in Amyloid Precursor Protein/Presenilin 1(APP/PS1)double transgenic mice and its possible mechani...Objective To explore whether the protein Deglycase protein 1(DJ1)can ameliorate Alzheimer’s disease(AD)-like pathology in Amyloid Precursor Protein/Presenilin 1(APP/PS1)double transgenic mice and its possible mechanism to provide a theoretical basis for exploring the pathogenesis of AD.Methods Adeno-associated viral vectors(AAV)of DJ1-overexpression or DJ1-knockdown were injected into the hippocampus of 7-month-old APP/PS1 mice to construct models of overexpression or knockdown.Mice were divided into the AD model control group(MC),AAV vector control group(NC),DJ1-overexpression group(DJ1+),and DJ1-knockdown group(DJ1-).After 21 days,the Morris water maze test,immunohistochemistry,immunofluorescence,and western blotting were used to evaluate the effects of DJ1 on mice.Results DJ1+overexpression decreased the latency and increased the number of platform traversals in the water maze test.DJ1-cells were cured and atrophied,and the intercellular structure was relaxed;the number of age spots and the expression of AD-related proteins were significantly increased.DJ1+increased the protein expression of Nuclear factor erythroid 2-related factor 2(NRF2),heme oxygenase-1(HO-1),light chain 3(LC3),phosphorylated AMPK(p-AMPK),and B cell lymphoma-2(BCL-2),as well as the antioxidant levels of total superoxide dismutase(T-SOD),total antioxidant capacity(T-AOC),and Glutathione peroxidase(GSH-PX),while decreasing the levels of Kelch-like hydrates-associated protein 1(Keap1),mammalian target of rapamycin(mTOR),p62/sequestosome1(p62/SQSTM1),Caspase3,and malondialdehyde(MDA).Conclusion DJ1-overexpression can ameliorate learning,memory,and AD-like pathology in APP/PS1 mice,which may be related to the activation of the NRF2/HO-1 and AMPK/mTOR pathways by DJ1.展开更多
文摘以编目分类和规则匹配为主的古籍文本主题分类方法存在工作效能低、专家知识依赖性强、分类依据单一化、古籍文本主题自动分类难等问题。对此,本文结合古籍文本内容和文字特征,尝试从古籍内容分类得到符合研究者需求的主题,推动数字人文研究范式的转型。首先,参照东汉古籍《说文解字》对文字的分析方式,以前期标注的古籍语料数据集为基础,构建全新的“字音(说)-原文(文)-结构(解)-字形(字)”四维特征数据集。其次,设计四维特征向量提取模型(speaking,word,pattern,and font to vector,SWPF2vec),并结合预训练模型实现对古籍文本细粒度的特征表示。再其次,构建融合卷积神经网络、循环神经网络和多头注意力机制的古籍文本主题分类模型(dianji-recurrent convolutional neural networks for text classification,DJ-TextRCNN)。最后,融入四维语义特征,实现对古籍文本多维度、深层次、细粒度的语义挖掘。在古籍文本主题分类任务上,DJ-TextRCNN模型在不同维度特征下的主题分类准确率均为最优,在“说文解字”四维特征下达到76.23%的准确率,初步实现了对古籍文本的精准主题分类。
文摘传统编目分类和规则匹配方法存在工作效能低、过度依赖专家知识、缺乏对古籍文本自身语义的深层次挖掘、编目主题边界模糊、较难实现对古籍文本领域主题的精准推荐等问题。为此,本文结合古籍语料特征探究如何实现精准推荐符合研究者需求的文本主题内容的方法,以推动数字人文研究的进一步发展。首先,选取本课题组前期标注的古籍语料数据进行主题类别标注和视图分类;其次,构建融合BERT(bidirectional encoder representation from transformers)预训练模型、改进卷积神经网络、循环神经网络和多头注意力机制的语义挖掘模型;最后,融入“主体-关系-客体”多视图的语义增强模型,构建DJ-TextRCNN(DianJi-recurrent convolutional neural networks for text classification)模型实现对典籍文本更细粒度、更深层次、更多维度的语义挖掘。研究结果发现,DJ-TextRCNN模型在不同视图下的古籍主题推荐任务的准确率均为最优。在“主体-关系-客体”视图下,精确率达到88.54%,初步实现了对古籍文本的精准主题推荐,对中华文化深层次、细粒度的语义挖掘具有一定的指导意义。
基金supported by the National Natural Science Foundation of China[grant numbers 81872626 and 82003454]Chinese Nutrition Society-Bright Moon Seaweed Group Nutrition and Health Research Fund[grant number CNS-BMSG2020A63]Key R&D and promotion projects in Henan Province[grant number 212102310219 and 212102310110]。
文摘Objective To explore whether the protein Deglycase protein 1(DJ1)can ameliorate Alzheimer’s disease(AD)-like pathology in Amyloid Precursor Protein/Presenilin 1(APP/PS1)double transgenic mice and its possible mechanism to provide a theoretical basis for exploring the pathogenesis of AD.Methods Adeno-associated viral vectors(AAV)of DJ1-overexpression or DJ1-knockdown were injected into the hippocampus of 7-month-old APP/PS1 mice to construct models of overexpression or knockdown.Mice were divided into the AD model control group(MC),AAV vector control group(NC),DJ1-overexpression group(DJ1+),and DJ1-knockdown group(DJ1-).After 21 days,the Morris water maze test,immunohistochemistry,immunofluorescence,and western blotting were used to evaluate the effects of DJ1 on mice.Results DJ1+overexpression decreased the latency and increased the number of platform traversals in the water maze test.DJ1-cells were cured and atrophied,and the intercellular structure was relaxed;the number of age spots and the expression of AD-related proteins were significantly increased.DJ1+increased the protein expression of Nuclear factor erythroid 2-related factor 2(NRF2),heme oxygenase-1(HO-1),light chain 3(LC3),phosphorylated AMPK(p-AMPK),and B cell lymphoma-2(BCL-2),as well as the antioxidant levels of total superoxide dismutase(T-SOD),total antioxidant capacity(T-AOC),and Glutathione peroxidase(GSH-PX),while decreasing the levels of Kelch-like hydrates-associated protein 1(Keap1),mammalian target of rapamycin(mTOR),p62/sequestosome1(p62/SQSTM1),Caspase3,and malondialdehyde(MDA).Conclusion DJ1-overexpression can ameliorate learning,memory,and AD-like pathology in APP/PS1 mice,which may be related to the activation of the NRF2/HO-1 and AMPK/mTOR pathways by DJ1.
