语音识别中的一个重要的分支就是关键词检索。虽然在英语上的关键词检索已经成熟,但是低资源的语音,比如维语的语音关键词检索研究缓慢,仍需要更深入的研究。文章在维吾尔语语数据集thuyg20上,先在GMM-HMM(Gaussian Mixture Model Hidde...语音识别中的一个重要的分支就是关键词检索。虽然在英语上的关键词检索已经成熟,但是低资源的语音,比如维语的语音关键词检索研究缓慢,仍需要更深入的研究。文章在维吾尔语语数据集thuyg20上,先在GMM-HMM(Gaussian Mixture Model Hidden Markov Model)声学模型,DNN-HMM(Hidden Markov Model Deep Neural Network)声学模型,LSTM-HMM(Long Short-term Memory Hidden Markov Model)声学模型解码产生的网格lattice上捕捉关键词,将DNN-HMM和LSTM-HMM解码产生的网格进行融合,再在融合的网格lattice上进行关键词检索。实验结果表明,融合后的结果在准确率和召回率方面要优于DNN-HMM和LSTM-HMM模型的检索性能。展开更多
老挝语属于低资源语言,在自然语言处理方面的研究较为薄弱,针对相对复杂的实体关系抽取技术暂时还没有相关研究.因此,该文通过研究老挝语的语言特点,提出了一种融合词性位置的无监督老挝语实体关系抽取方法.首先,通过爬虫技术在中国国...老挝语属于低资源语言,在自然语言处理方面的研究较为薄弱,针对相对复杂的实体关系抽取技术暂时还没有相关研究.因此,该文通过研究老挝语的语言特点,提出了一种融合词性位置的无监督老挝语实体关系抽取方法.首先,通过爬虫技术在中国国际广播电台网(China Radio International,CRI)上爬取老挝语新闻3000篇,其次,对爬取的新闻文本进行预处理,包括分句,分词,命名实体识别以及词性标注;然后,使用融合了距离特征、TF-IDF特征和词性位置特征的词的分布式表示上下文特征,将其作为聚类模型的输入;之后,使用改进的KANN-DBSCAN聚类模型在预处理后的语料集上进行聚类试验;最后,采用优化的DCM方法对聚类结果进行关系标注,选出簇的关系标签,最终得到关系三元组.实验结果表明,该文提出的方法在低资源语言老挝语上达到了不错的效果,平均准确率达到了60.43%.展开更多
Objective: Pelvic organ prolapse is an emerging public health problem affecting adult women of all ages with a negative impact on social, physical well-being, and psychological. Its presents several challenges in coun...Objective: Pelvic organ prolapse is an emerging public health problem affecting adult women of all ages with a negative impact on social, physical well-being, and psychological. Its presents several challenges in countries with low resources. This literature review aims to examine POP in its epidemiological aspects, risk factors, and staging by taking up the challenges associated with low-resource settings and identifying some avenues for future research. Methods: We searched the PubMed, Google Scholar, and Scopus databases. The other studies were identified by checking the secondary references in the original citation. We have collected studies on adult women published in English for the last 30 years. In total, 71 articles were read. We excluded studies from all newspaper articles, Studies presenting co-morbidities (fistulas, cervical cancer, pregnancy), those evaluating treatment, letters, comments, case reports, practice guidelines, news, historical articles, legal cases, published erratum, and congresses. Results: 16 studies examining the epidemiology have been identified with 11 in countries defined by the World Bank as limited or intermediate resources. 18 on risk factors whose 10 in countries with limited or intermediate resources, 10 on staging and 27 on physiopathology. Conclusion: POP affects the young more in low-resource settings. Its prevalence remains underestimated for several reasons. Several risk factors found are the same as those of women in countries with a high standard of living. However, there are some specific risk factors for these resource-limited settings.展开更多
Coronavirus disease (COVID-19) is an infectious disease caused by a newly discovered coronavirus. This novel coronavirus is called SARS-COV-2 and the disease that it causes is called COVID-19 causes serious respirator...Coronavirus disease (COVID-19) is an infectious disease caused by a newly discovered coronavirus. This novel coronavirus is called SARS-COV-2 and the disease that it causes is called COVID-19 causes serious respiratory morbidity and mortality. We aim to spot COVID-19 disease during pregnancy during this pandemic era. COVID-19 has various clinical presentations almost same in pregnant and non-pregnant victims. The hallmark for treatment is supportive management. Vertical transmission is probable but still not well confirmed and no evidence for virus in amniotic fluid, cord blood, neonatal throat swabs, placenta swabs, genital fluid and breastmilk samples from COVID-19 infected mothers. Research is running everywhere looking step ahead for actual treatment and vaccination. COVID-19 is newly emerged disease still not well explored in many aspects looking for soon definite treatment and vaccination for prevention to roll COVID-19.展开更多
知识库问答(Knowledge Base Question Answering,KBQA)借助知识库中精度高、关联性强的结构化知识,为给定的复杂事实型问句提供准确、简短的答案.语义解析是知识库问答的主流方法之一,该类方法在给定的问句语义表征形式下,将非结构化的...知识库问答(Knowledge Base Question Answering,KBQA)借助知识库中精度高、关联性强的结构化知识,为给定的复杂事实型问句提供准确、简短的答案.语义解析是知识库问答的主流方法之一,该类方法在给定的问句语义表征形式下,将非结构化的问句映射为结构化的语义表征,再将其改写为知识库查询获取答案.目前,面向知识库问答的语义解析方法主要面临三个挑战:首先是如何选择合适的语义表征形式以表达问句的语义,然后是如何解析问句的复杂语义并输出相应的语义表征,最后是如何应对特定领域中数据标注成本高昂、高质量数据匮乏的问题.本文从上述挑战出发,分析了知识库问答中常用的语义表征的特点与不足,然后梳理现有方法并总结分析其如何应对问句的复杂语义,接着介绍了当前方法在标注数据匮乏的低资源场景下的尝试,最后展望并讨论了面向知识库问答的语义解析的未来发展方向.展开更多
文摘老挝语属于低资源语言,在自然语言处理方面的研究较为薄弱,针对相对复杂的实体关系抽取技术暂时还没有相关研究.因此,该文通过研究老挝语的语言特点,提出了一种融合词性位置的无监督老挝语实体关系抽取方法.首先,通过爬虫技术在中国国际广播电台网(China Radio International,CRI)上爬取老挝语新闻3000篇,其次,对爬取的新闻文本进行预处理,包括分句,分词,命名实体识别以及词性标注;然后,使用融合了距离特征、TF-IDF特征和词性位置特征的词的分布式表示上下文特征,将其作为聚类模型的输入;之后,使用改进的KANN-DBSCAN聚类模型在预处理后的语料集上进行聚类试验;最后,采用优化的DCM方法对聚类结果进行关系标注,选出簇的关系标签,最终得到关系三元组.实验结果表明,该文提出的方法在低资源语言老挝语上达到了不错的效果,平均准确率达到了60.43%.
文摘Objective: Pelvic organ prolapse is an emerging public health problem affecting adult women of all ages with a negative impact on social, physical well-being, and psychological. Its presents several challenges in countries with low resources. This literature review aims to examine POP in its epidemiological aspects, risk factors, and staging by taking up the challenges associated with low-resource settings and identifying some avenues for future research. Methods: We searched the PubMed, Google Scholar, and Scopus databases. The other studies were identified by checking the secondary references in the original citation. We have collected studies on adult women published in English for the last 30 years. In total, 71 articles were read. We excluded studies from all newspaper articles, Studies presenting co-morbidities (fistulas, cervical cancer, pregnancy), those evaluating treatment, letters, comments, case reports, practice guidelines, news, historical articles, legal cases, published erratum, and congresses. Results: 16 studies examining the epidemiology have been identified with 11 in countries defined by the World Bank as limited or intermediate resources. 18 on risk factors whose 10 in countries with limited or intermediate resources, 10 on staging and 27 on physiopathology. Conclusion: POP affects the young more in low-resource settings. Its prevalence remains underestimated for several reasons. Several risk factors found are the same as those of women in countries with a high standard of living. However, there are some specific risk factors for these resource-limited settings.
文摘Coronavirus disease (COVID-19) is an infectious disease caused by a newly discovered coronavirus. This novel coronavirus is called SARS-COV-2 and the disease that it causes is called COVID-19 causes serious respiratory morbidity and mortality. We aim to spot COVID-19 disease during pregnancy during this pandemic era. COVID-19 has various clinical presentations almost same in pregnant and non-pregnant victims. The hallmark for treatment is supportive management. Vertical transmission is probable but still not well confirmed and no evidence for virus in amniotic fluid, cord blood, neonatal throat swabs, placenta swabs, genital fluid and breastmilk samples from COVID-19 infected mothers. Research is running everywhere looking step ahead for actual treatment and vaccination. COVID-19 is newly emerged disease still not well explored in many aspects looking for soon definite treatment and vaccination for prevention to roll COVID-19.
文摘知识库问答(Knowledge Base Question Answering,KBQA)借助知识库中精度高、关联性强的结构化知识,为给定的复杂事实型问句提供准确、简短的答案.语义解析是知识库问答的主流方法之一,该类方法在给定的问句语义表征形式下,将非结构化的问句映射为结构化的语义表征,再将其改写为知识库查询获取答案.目前,面向知识库问答的语义解析方法主要面临三个挑战:首先是如何选择合适的语义表征形式以表达问句的语义,然后是如何解析问句的复杂语义并输出相应的语义表征,最后是如何应对特定领域中数据标注成本高昂、高质量数据匮乏的问题.本文从上述挑战出发,分析了知识库问答中常用的语义表征的特点与不足,然后梳理现有方法并总结分析其如何应对问句的复杂语义,接着介绍了当前方法在标注数据匮乏的低资源场景下的尝试,最后展望并讨论了面向知识库问答的语义解析的未来发展方向.
