组织机构名识别是命名实体识别的核心任务之一,也是最困难的任务。近年来,预训练模型在中文自然语言处理领域得到广泛应用,预训练的词嵌入模型在中文命名实体识别上取得了非常好的效果,但是在组织机构名识别上还有很大的提升空间。针对...组织机构名识别是命名实体识别的核心任务之一,也是最困难的任务。近年来,预训练模型在中文自然语言处理领域得到广泛应用,预训练的词嵌入模型在中文命名实体识别上取得了非常好的效果,但是在组织机构名识别上还有很大的提升空间。针对这一问题,改进ELMO(embedding from language models)预训练模型,结合双向LSTM神经网络模型和条件随机场模型,去识别组织机构名。对于ELMO的改进,主要通过筛选高频机构词,然后将高频机构词加入中文字典,通过ELMO模型训练生成机构词向量和普通字向量。字向量不用考虑未登录词的问题,机构词向量引入了先验知识,结合起来可以使得生成的字词向量能够更好地表征组织机构名。实验结果表明,预训练模型的数据集相对较小时,该方法比字向量嵌入的方法有更好的效果,F1值提高了1.3%。展开更多
[Objective] To examine the grammar model based on lexical substring exac- tion for RNA secondary structure prediction. [Method] By introducing cloud model into stochastic grammar model, a machine learning algorithm su...[Objective] To examine the grammar model based on lexical substring exac- tion for RNA secondary structure prediction. [Method] By introducing cloud model into stochastic grammar model, a machine learning algorithm suitable for the lexicalized stochastic grammar model was proposed. The word grid mode was used to extract and divide RNA sequence to acquire lexical substring, and the cloud classifier was used to search the maximum probability of each lemma which was marked as a certain sec- ondary structure type. Then, the lemma information was introduced into the training stochastic grammar process as prior information, realizing the prediction on the sec- ondary structure of RNA, and the method was tested by experiment. [Result] The experimental results showed that the prediction accuracy and searching speed of stochastic grammar cloud model were significantly improved from the prediction with simple stochastic grammar. [Conclusion] This study laid the foundation for the wide application of stochastic grammar model for RNA secondary structure prediction.展开更多
Learning keywords is one of the best ways to keep abreast of the latest developments in a country.The China Academy of Translation,a research institute affiliated with the China International Publishing Group,the coun...Learning keywords is one of the best ways to keep abreast of the latest developments in a country.The China Academy of Translation,a research institute affiliated with the China International Publishing Group,the country’s leading international publisher,regularly analyzes prevailing Chinese terms in various展开更多
Learning keywords is one of the best ways to keep abreast of the latest developments in a country.The China Academy of Translation,a research institute affiliated with the China International Publishing Group,the coun...Learning keywords is one of the best ways to keep abreast of the latest developments in a country.The China Academy of Translation,a research institute affiliated with the China International Publishing Group,the country’s leading international publisher,regularly analyzes prevailing Chinese terms in various sectors and translates them into a number of foreign languages ranging from English to Arabic.展开更多
文摘组织机构名识别是命名实体识别的核心任务之一,也是最困难的任务。近年来,预训练模型在中文自然语言处理领域得到广泛应用,预训练的词嵌入模型在中文命名实体识别上取得了非常好的效果,但是在组织机构名识别上还有很大的提升空间。针对这一问题,改进ELMO(embedding from language models)预训练模型,结合双向LSTM神经网络模型和条件随机场模型,去识别组织机构名。对于ELMO的改进,主要通过筛选高频机构词,然后将高频机构词加入中文字典,通过ELMO模型训练生成机构词向量和普通字向量。字向量不用考虑未登录词的问题,机构词向量引入了先验知识,结合起来可以使得生成的字词向量能够更好地表征组织机构名。实验结果表明,预训练模型的数据集相对较小时,该方法比字向量嵌入的方法有更好的效果,F1值提高了1.3%。
基金Supported by the Science Foundation of Hengyang Normal University of China(09A36)~~
文摘[Objective] To examine the grammar model based on lexical substring exac- tion for RNA secondary structure prediction. [Method] By introducing cloud model into stochastic grammar model, a machine learning algorithm suitable for the lexicalized stochastic grammar model was proposed. The word grid mode was used to extract and divide RNA sequence to acquire lexical substring, and the cloud classifier was used to search the maximum probability of each lemma which was marked as a certain sec- ondary structure type. Then, the lemma information was introduced into the training stochastic grammar process as prior information, realizing the prediction on the sec- ondary structure of RNA, and the method was tested by experiment. [Result] The experimental results showed that the prediction accuracy and searching speed of stochastic grammar cloud model were significantly improved from the prediction with simple stochastic grammar. [Conclusion] This study laid the foundation for the wide application of stochastic grammar model for RNA secondary structure prediction.
文摘Learning keywords is one of the best ways to keep abreast of the latest developments in a country.The China Academy of Translation,a research institute affiliated with the China International Publishing Group,the country’s leading international publisher,regularly analyzes prevailing Chinese terms in various
文摘Learning keywords is one of the best ways to keep abreast of the latest developments in a country.The China Academy of Translation,a research institute affiliated with the China International Publishing Group,the country’s leading international publisher,regularly analyzes prevailing Chinese terms in various sectors and translates them into a number of foreign languages ranging from English to Arabic.
