In the process of online English learning,learning efficiency is influenced by various non-intelligence factors.Non-intelligence factors include learning motivation,self-efficacy,cultural background knowledge and onli...In the process of online English learning,learning efficiency is influenced by various non-intelligence factors.Non-intelligence factors include learning motivation,self-efficacy,cultural background knowledge and online learning strategies which play crucial roles in online English learning.This paper analyses the non-intelligence factors affecting students'online English learning and put forwards some measures to enhance online English learning.展开更多
Demotivation can be defined as"specific external forces that reduce or diminish the motivational basis of a behavioral intention or an ongoing action".The purpose of the present study is to identify what fac...Demotivation can be defined as"specific external forces that reduce or diminish the motivational basis of a behavioral intention or an ongoing action".The purpose of the present study is to identify what factors demotivated non-English majors in a southeast university in China.Questionnaire and semi-interview were developed to collect data.82 Participants responded to the questionnaire,three of whom shared their perception and experience of English learning in the interview.Consequently,this study identified four salient demotivating factors which made non-English majors less enthusiastic about English learning:(1)unhappy and unsuccessful experience of English learning in the elementary and high school;(2)low language proficiency;(3)difficulties of language learning contents at university and(4)utilitarian ideas.It is hoped that administration and English teachers will match the course,teaching content and course book with students’need,interest and proficiency.展开更多
触发执行编程(Trigger-Action Programming,TAP)为用户联动物联网(Internet of Things,IoT)设备提供了便捷的编程范式。利用机器学习对用户已编辑的TAP规则进行分析,实现TAP规则推荐和生成等功能可以提升用户体验。但TAP规则可能包含个...触发执行编程(Trigger-Action Programming,TAP)为用户联动物联网(Internet of Things,IoT)设备提供了便捷的编程范式。利用机器学习对用户已编辑的TAP规则进行分析,实现TAP规则推荐和生成等功能可以提升用户体验。但TAP规则可能包含个人隐私信息,用户对上传和分享TAP信息存在顾虑。文章提出了基于联邦学习和区块链技术的TAP规则处理系统,用户可在本地进行TAP模型训练,无需上传隐私数据。为解决集中式服务器单点故障和防范恶意模型参数上传的问题,文章利用区块链技术改进集中式TAP联邦学习架构。用户将本地模型更新的累积梯度传输给区块链中的矿工,进行异常识别和交叉验证。矿工委员会整合正常用户提供的累积梯度,得到的全局模型作为一个新区块的数据,链接到区块链上,供用户下载使用。文章采用轻量级无监督的非负矩阵分解方法验证了提出的基于联邦学习和区块链的分布式学习架构的有效性。实验证明该联邦学习架构能有效保护TAP数据中的隐私,并且区块链中的矿工能够很好地识别恶意模型参数,确保了模型的稳定性。展开更多
无监督特征选择是机器学习和数据挖掘中的一种重要的降维技术。然而当前的无监督特征选择方法侧重于从数据的邻接矩阵中学习数据的流形结构,忽视非邻接数据对之间的关联。其次这些方法都假设数据实例具有独立同一性,但现实中的数据样本...无监督特征选择是机器学习和数据挖掘中的一种重要的降维技术。然而当前的无监督特征选择方法侧重于从数据的邻接矩阵中学习数据的流形结构,忽视非邻接数据对之间的关联。其次这些方法都假设数据实例具有独立同一性,但现实中的数据样本其来源是不同的,这样的假设就不成立。此外,在原始数据空间中特征重要性的衡量会受到数据和特征中的噪声影响。基于以上问题,本文提出了潜在多步马尔可夫概率的鲁棒无监督特征选择方法(unsupervised feature selection via multi-step Markov probability and latent representation,MMLRL),其思想是通过最大多步马尔可夫转移概率学习数据流形结构,然后通过对称非负矩阵分解模型学习数据的潜在表示,最后在数据的潜在表示空间中选择特征。同时在6个不同类型的数据集上验证了所提出算法的有效性。展开更多
文摘In the process of online English learning,learning efficiency is influenced by various non-intelligence factors.Non-intelligence factors include learning motivation,self-efficacy,cultural background knowledge and online learning strategies which play crucial roles in online English learning.This paper analyses the non-intelligence factors affecting students'online English learning and put forwards some measures to enhance online English learning.
文摘Demotivation can be defined as"specific external forces that reduce or diminish the motivational basis of a behavioral intention or an ongoing action".The purpose of the present study is to identify what factors demotivated non-English majors in a southeast university in China.Questionnaire and semi-interview were developed to collect data.82 Participants responded to the questionnaire,three of whom shared their perception and experience of English learning in the interview.Consequently,this study identified four salient demotivating factors which made non-English majors less enthusiastic about English learning:(1)unhappy and unsuccessful experience of English learning in the elementary and high school;(2)low language proficiency;(3)difficulties of language learning contents at university and(4)utilitarian ideas.It is hoped that administration and English teachers will match the course,teaching content and course book with students’need,interest and proficiency.
文摘触发执行编程(Trigger-Action Programming,TAP)为用户联动物联网(Internet of Things,IoT)设备提供了便捷的编程范式。利用机器学习对用户已编辑的TAP规则进行分析,实现TAP规则推荐和生成等功能可以提升用户体验。但TAP规则可能包含个人隐私信息,用户对上传和分享TAP信息存在顾虑。文章提出了基于联邦学习和区块链技术的TAP规则处理系统,用户可在本地进行TAP模型训练,无需上传隐私数据。为解决集中式服务器单点故障和防范恶意模型参数上传的问题,文章利用区块链技术改进集中式TAP联邦学习架构。用户将本地模型更新的累积梯度传输给区块链中的矿工,进行异常识别和交叉验证。矿工委员会整合正常用户提供的累积梯度,得到的全局模型作为一个新区块的数据,链接到区块链上,供用户下载使用。文章采用轻量级无监督的非负矩阵分解方法验证了提出的基于联邦学习和区块链的分布式学习架构的有效性。实验证明该联邦学习架构能有效保护TAP数据中的隐私,并且区块链中的矿工能够很好地识别恶意模型参数,确保了模型的稳定性。
文摘无监督特征选择是机器学习和数据挖掘中的一种重要的降维技术。然而当前的无监督特征选择方法侧重于从数据的邻接矩阵中学习数据的流形结构,忽视非邻接数据对之间的关联。其次这些方法都假设数据实例具有独立同一性,但现实中的数据样本其来源是不同的,这样的假设就不成立。此外,在原始数据空间中特征重要性的衡量会受到数据和特征中的噪声影响。基于以上问题,本文提出了潜在多步马尔可夫概率的鲁棒无监督特征选择方法(unsupervised feature selection via multi-step Markov probability and latent representation,MMLRL),其思想是通过最大多步马尔可夫转移概率学习数据流形结构,然后通过对称非负矩阵分解模型学习数据的潜在表示,最后在数据的潜在表示空间中选择特征。同时在6个不同类型的数据集上验证了所提出算法的有效性。
