列控车载设备是保障高速列车行车安全、提高运输效率的核心组成部分,快速有效地诊断其故障类型具有重要意义.针对300T型列控车载设备故障文本数据的错综性和时序性,提出一种基于LSTM-BP级联网络模型的车载设备智能故障诊断方法.首先,采...列控车载设备是保障高速列车行车安全、提高运输效率的核心组成部分,快速有效地诊断其故障类型具有重要意义.针对300T型列控车载设备故障文本数据的错综性和时序性,提出一种基于LSTM-BP级联网络模型的车载设备智能故障诊断方法.首先,采用贝叶斯正则化(Bayesian Regularization,BR)算法优化BP神经网络提高模型的泛化能力;其次,利用长短时记忆网络(Long Short Term Memory Network,LSTM)的记忆特性,充分学习具有时序性的故障特征信息,解决BP神经网络模型难以准确诊断关机误报和引发故障等问题;最后,利用实际数据对模型进行多次试验分析,BR优化的神经网络模型分类准确率为85.06%;而LSTM-BP级联网络模型分类准确率达到95.10%,能够很好地解决对关机误报和引发故障诊断不准确的问题,验证了本文所提出的智能故障诊断方法的有效性.展开更多
This paper examines ceramic-related cultural texts as a case study,systematically evaluating the capabilities and limitations of two popular large language models(LLMs)when processing culturally embedded content while...This paper examines ceramic-related cultural texts as a case study,systematically evaluating the capabilities and limitations of two popular large language models(LLMs)when processing culturally embedded content while simultaneously developing innovative methodological approaches for technology-enhanced translation classrooms.By conducting comparative analyses of artificial intelligence(AI)-generated translations,the study identifies key challenges in translating ceramic cultural texts,explores potential refinements for machine translation algorithms,and formulates evidence-based teaching strategies that leverage these insights to cultivate comprehensive translation skills.The findings indicate that while LLMs have demonstrated notable effectiveness in basic information transfer and literal semantic comprehension,they currently still need improvements to understand and process specialized jargon as well as metaphors.The findings also offer translation teachers a substantive framework for pedagogical transformation in the digital era,effectively bridging the theoretical divide between cultural translation studies and technological applications in translation education.AI should be leveraged to enhance ceramic culture translation,facilitating the advancement of cross-cultural communication and translation strategies.展开更多
文摘列控车载设备是保障高速列车行车安全、提高运输效率的核心组成部分,快速有效地诊断其故障类型具有重要意义.针对300T型列控车载设备故障文本数据的错综性和时序性,提出一种基于LSTM-BP级联网络模型的车载设备智能故障诊断方法.首先,采用贝叶斯正则化(Bayesian Regularization,BR)算法优化BP神经网络提高模型的泛化能力;其次,利用长短时记忆网络(Long Short Term Memory Network,LSTM)的记忆特性,充分学习具有时序性的故障特征信息,解决BP神经网络模型难以准确诊断关机误报和引发故障等问题;最后,利用实际数据对模型进行多次试验分析,BR优化的神经网络模型分类准确率为85.06%;而LSTM-BP级联网络模型分类准确率达到95.10%,能够很好地解决对关机误报和引发故障诊断不准确的问题,验证了本文所提出的智能故障诊断方法的有效性.
文摘This paper examines ceramic-related cultural texts as a case study,systematically evaluating the capabilities and limitations of two popular large language models(LLMs)when processing culturally embedded content while simultaneously developing innovative methodological approaches for technology-enhanced translation classrooms.By conducting comparative analyses of artificial intelligence(AI)-generated translations,the study identifies key challenges in translating ceramic cultural texts,explores potential refinements for machine translation algorithms,and formulates evidence-based teaching strategies that leverage these insights to cultivate comprehensive translation skills.The findings indicate that while LLMs have demonstrated notable effectiveness in basic information transfer and literal semantic comprehension,they currently still need improvements to understand and process specialized jargon as well as metaphors.The findings also offer translation teachers a substantive framework for pedagogical transformation in the digital era,effectively bridging the theoretical divide between cultural translation studies and technological applications in translation education.AI should be leveraged to enhance ceramic culture translation,facilitating the advancement of cross-cultural communication and translation strategies.
