Speech Emotion Recognition(SER)has received widespread attention as a crucial way for understanding human emotional states.However,the impact of irrelevant information on speech signals and data sparsity limit the dev...Speech Emotion Recognition(SER)has received widespread attention as a crucial way for understanding human emotional states.However,the impact of irrelevant information on speech signals and data sparsity limit the development of SER system.To address these issues,this paper proposes a framework that incorporates the Attentive Mask Residual Network(AM-ResNet)and the self-supervised learning model Wav2vec 2.0 to obtain AM-ResNet features and Wav2vec 2.0 features respectively,together with a cross-attention module to interact and fuse these two features.The AM-ResNet branch mainly consists of maximum amplitude difference detection,mask residual block,and an attention mechanism.Among them,the maximum amplitude difference detection and the mask residual block act on the pre-processing and the network,respectively,to reduce the impact of silent frames,and the attention mechanism assigns different weights to unvoiced and voiced speech to reduce redundant emotional information caused by unvoiced speech.In the Wav2vec 2.0 branch,this model is introduced as a feature extractor to obtain general speech features(Wav2vec 2.0 features)through pre-training with a large amount of unlabeled speech data,which can assist the SER task and cope with data sparsity problems.In the cross-attention module,AM-ResNet features and Wav2vec 2.0 features are interacted with and fused to obtain the cross-fused features,which are used to predict the final emotion.Furthermore,multi-label learning is also used to add ambiguous emotion utterances to deal with data limitations.Finally,experimental results illustrate the usefulness and superiority of our proposed framework over existing state-of-the-art approaches.展开更多
为了深入研究我国慢性病医防融合领域的发展趋势和演化过程,本文收集了2006~2024年的373篇相关文献,经过数据清洗和预处理后,引入Word2vec的LDA模型进行文献的主题挖掘,确定每个时期的最佳主题数量,并生成主题演化桑基图。计算不同时间...为了深入研究我国慢性病医防融合领域的发展趋势和演化过程,本文收集了2006~2024年的373篇相关文献,经过数据清洗和预处理后,引入Word2vec的LDA模型进行文献的主题挖掘,确定每个时期的最佳主题数量,并生成主题演化桑基图。计算不同时间段内各主题强度,并通过交互式条形图描述热点主题。结果显示,在第一阶段2006~2020年,大部分研究主要集中在如何整合医疗服务,以及如何将慢性病防控与医防结合;在第二阶段2021~2022年,除了延续既有的主题,部分研究焦点转移到如何更好地管理和融合综合医疗服务,以及如何将公共卫生服务与医疗体系更有效地结合;在第三阶段2023~2024年,研究重点在于如何实现健康服务与医防的深度融合,以及如何在医疗服务中具体落实医防融合的理念,研究更加注重实际操作和具体应用。通过主题演化分析揭示了不同时期内主题之间的关联和演化过程,综合医疗服务、慢性病防控与医防结合等主题在不同阶段都有较强的延续性,而研究重点随着时间的推移逐渐从综合医疗服务向医防融合和健康服务管理方向转移。研究发现,一些主题在不同时期内保持较高的强度,从本研究主题强度图可以看出,在慢性病医防融合领域,社区基层医疗机构在医防融合中具有重要作用,此外2021年及以后的阶段中公共卫生体系建设及医防融合成为研究的共识热点。该研究有助于更全面地理解慢性病医防融合领域的研究动态,为未来的研究方向和政策制定提供有益的参考,同时也为文本分析方法的应用提供了实践示范。未来的研究可以进一步挖掘基层医疗与医防协同机制以及健康服务管理与慢性病防控方面的潜力,更好地帮助社区基层医疗机构服务提供者应对来自人口老龄化社会慢性病高发以及多样化健康需求的挑战,同时也要关注对应的新兴技术如人工智能和大数据分析和对应的数据隐私和伦理挑战,以及政策实施中的风险。In this paper, in order to deeply study the development trend and evolution process in the field of chronic disease medical preventive integration in China, 373 relevant literatures from 2006~2024 were collected, and after data cleaning and pre-processing, the LDA model of Word2vec was introduced in the theme mining of the literature to determine the optimal number of themes in each period and generate the theme evolution Sankey diagram. The intensity of each topic in different time periods is calculated and hot topics are described by interactive bar charts. The results show that in the first period of 2006~2020, most of the studies focused on how to integrate healthcare services and how to combine chronic disease prevention and control with medical prevention;in the second period of 2021~2022, in addition to the continuation of the existing themes, some of the studies shifted their focus to how to better manage and integrate integrated healthcare services and how to combine public health services with the healthcare system more effectively;in the third stage, 2023~2024, the research focused on how to realize the deep integration of health services and medical preventive, and how to implement the concept of medical prevention integration in health care services, and the research focused more on practical operation and specific application. The analysis of theme evolution reveals the connection and evolution process between themes in different periods. The themes of comprehensive medical service, chronic disease prevention and control and medical prevention integration have strong continuity in different stages, while the focus of research gradually shifts from comprehensive medical service to medical prevention integration and health service management over time. It is found that some themes maintain a high intensity in different periods, and the intensity map of the themes in this study shows that in the field of chronic disease medical prevention integration, community-based primary healthcare organizations have an important role in medical prevention integration, and in addition, public health system construction and medical prevention integration have become consensus hotspots in research in the stage of 2021 and beyond. This study contributes to a more comprehensive understanding of the research dynamics in the field of chronic disease medical prevention integration, provides useful references for future research directions and policy formulation, and also provides a practical demonstration of the application of text analysis methods. Future research can further explore the potential of primary care and medical prevention synergistic mechanisms as well as health service management and chronic disease prevention and control to better help community-based primary care providers to cope with the challenges from the high prevalence of chronic diseases and diversified health needs of an aging population, as well as to pay attention to the corresponding emerging technologies such as artificial intelligence and big data analytics and the corresponding data privacy and ethical challenges, and the risks in policy implementation.展开更多
本文结合了Node2vec和GCN这两种方法,先利用Node2vec方法得到初步的图嵌入,之后将其作为输入利用GCN进一步更新图嵌入矩阵。本文选择在维基数据集上进行节点分类任务,比较了结合前后方法的表现,验证了其有效性。In this paper, we integ...本文结合了Node2vec和GCN这两种方法,先利用Node2vec方法得到初步的图嵌入,之后将其作为输入利用GCN进一步更新图嵌入矩阵。本文选择在维基数据集上进行节点分类任务,比较了结合前后方法的表现,验证了其有效性。In this paper, we integrate the Node2vec and GCN methods. Initially, the Node2vec method is employed to obtain preliminary graph embeddings, which are then used as input to further update the graph embedding matrix through GCN. The study selects the Wikipedia dataset for node classification tasks, comparing the performance of the methods before and after integration to validate their effectiveness.展开更多
基金supported by Chongqing University of Posts and Telecommunications Ph.D.Innovative Talents Project(Grant No.BYJS202106)Chongqing Postgraduate Research Innovation Project(Grant No.CYB21203).
文摘Speech Emotion Recognition(SER)has received widespread attention as a crucial way for understanding human emotional states.However,the impact of irrelevant information on speech signals and data sparsity limit the development of SER system.To address these issues,this paper proposes a framework that incorporates the Attentive Mask Residual Network(AM-ResNet)and the self-supervised learning model Wav2vec 2.0 to obtain AM-ResNet features and Wav2vec 2.0 features respectively,together with a cross-attention module to interact and fuse these two features.The AM-ResNet branch mainly consists of maximum amplitude difference detection,mask residual block,and an attention mechanism.Among them,the maximum amplitude difference detection and the mask residual block act on the pre-processing and the network,respectively,to reduce the impact of silent frames,and the attention mechanism assigns different weights to unvoiced and voiced speech to reduce redundant emotional information caused by unvoiced speech.In the Wav2vec 2.0 branch,this model is introduced as a feature extractor to obtain general speech features(Wav2vec 2.0 features)through pre-training with a large amount of unlabeled speech data,which can assist the SER task and cope with data sparsity problems.In the cross-attention module,AM-ResNet features and Wav2vec 2.0 features are interacted with and fused to obtain the cross-fused features,which are used to predict the final emotion.Furthermore,multi-label learning is also used to add ambiguous emotion utterances to deal with data limitations.Finally,experimental results illustrate the usefulness and superiority of our proposed framework over existing state-of-the-art approaches.
文摘为了深入研究我国慢性病医防融合领域的发展趋势和演化过程,本文收集了2006~2024年的373篇相关文献,经过数据清洗和预处理后,引入Word2vec的LDA模型进行文献的主题挖掘,确定每个时期的最佳主题数量,并生成主题演化桑基图。计算不同时间段内各主题强度,并通过交互式条形图描述热点主题。结果显示,在第一阶段2006~2020年,大部分研究主要集中在如何整合医疗服务,以及如何将慢性病防控与医防结合;在第二阶段2021~2022年,除了延续既有的主题,部分研究焦点转移到如何更好地管理和融合综合医疗服务,以及如何将公共卫生服务与医疗体系更有效地结合;在第三阶段2023~2024年,研究重点在于如何实现健康服务与医防的深度融合,以及如何在医疗服务中具体落实医防融合的理念,研究更加注重实际操作和具体应用。通过主题演化分析揭示了不同时期内主题之间的关联和演化过程,综合医疗服务、慢性病防控与医防结合等主题在不同阶段都有较强的延续性,而研究重点随着时间的推移逐渐从综合医疗服务向医防融合和健康服务管理方向转移。研究发现,一些主题在不同时期内保持较高的强度,从本研究主题强度图可以看出,在慢性病医防融合领域,社区基层医疗机构在医防融合中具有重要作用,此外2021年及以后的阶段中公共卫生体系建设及医防融合成为研究的共识热点。该研究有助于更全面地理解慢性病医防融合领域的研究动态,为未来的研究方向和政策制定提供有益的参考,同时也为文本分析方法的应用提供了实践示范。未来的研究可以进一步挖掘基层医疗与医防协同机制以及健康服务管理与慢性病防控方面的潜力,更好地帮助社区基层医疗机构服务提供者应对来自人口老龄化社会慢性病高发以及多样化健康需求的挑战,同时也要关注对应的新兴技术如人工智能和大数据分析和对应的数据隐私和伦理挑战,以及政策实施中的风险。In this paper, in order to deeply study the development trend and evolution process in the field of chronic disease medical preventive integration in China, 373 relevant literatures from 2006~2024 were collected, and after data cleaning and pre-processing, the LDA model of Word2vec was introduced in the theme mining of the literature to determine the optimal number of themes in each period and generate the theme evolution Sankey diagram. The intensity of each topic in different time periods is calculated and hot topics are described by interactive bar charts. The results show that in the first period of 2006~2020, most of the studies focused on how to integrate healthcare services and how to combine chronic disease prevention and control with medical prevention;in the second period of 2021~2022, in addition to the continuation of the existing themes, some of the studies shifted their focus to how to better manage and integrate integrated healthcare services and how to combine public health services with the healthcare system more effectively;in the third stage, 2023~2024, the research focused on how to realize the deep integration of health services and medical preventive, and how to implement the concept of medical prevention integration in health care services, and the research focused more on practical operation and specific application. The analysis of theme evolution reveals the connection and evolution process between themes in different periods. The themes of comprehensive medical service, chronic disease prevention and control and medical prevention integration have strong continuity in different stages, while the focus of research gradually shifts from comprehensive medical service to medical prevention integration and health service management over time. It is found that some themes maintain a high intensity in different periods, and the intensity map of the themes in this study shows that in the field of chronic disease medical prevention integration, community-based primary healthcare organizations have an important role in medical prevention integration, and in addition, public health system construction and medical prevention integration have become consensus hotspots in research in the stage of 2021 and beyond. This study contributes to a more comprehensive understanding of the research dynamics in the field of chronic disease medical prevention integration, provides useful references for future research directions and policy formulation, and also provides a practical demonstration of the application of text analysis methods. Future research can further explore the potential of primary care and medical prevention synergistic mechanisms as well as health service management and chronic disease prevention and control to better help community-based primary care providers to cope with the challenges from the high prevalence of chronic diseases and diversified health needs of an aging population, as well as to pay attention to the corresponding emerging technologies such as artificial intelligence and big data analytics and the corresponding data privacy and ethical challenges, and the risks in policy implementation.
文摘本文结合了Node2vec和GCN这两种方法,先利用Node2vec方法得到初步的图嵌入,之后将其作为输入利用GCN进一步更新图嵌入矩阵。本文选择在维基数据集上进行节点分类任务,比较了结合前后方法的表现,验证了其有效性。In this paper, we integrate the Node2vec and GCN methods. Initially, the Node2vec method is employed to obtain preliminary graph embeddings, which are then used as input to further update the graph embedding matrix through GCN. The study selects the Wikipedia dataset for node classification tasks, comparing the performance of the methods before and after integration to validate their effectiveness.
