The advantages and disadvantages for learning English in the Network-based environment attract most researchers’concern nowadays.This study profiles college English teachers’beliefs about the networkbased language l...The advantages and disadvantages for learning English in the Network-based environment attract most researchers’concern nowadays.This study profiles college English teachers’beliefs about the networkbased language learning.The main finding is that teachers’beliefs about network-based language learning are heterogeneous and thus reflect a wide range in terms of the evolution of approaches and technology use.展开更多
Based on the previous research results and five principles,this paper tries to build an applicable optimization framew ork of college English classroom teaching,w hich centers on teacher development.According to this ...Based on the previous research results and five principles,this paper tries to build an applicable optimization framew ork of college English classroom teaching,w hich centers on teacher development.According to this framew ork,teacher development should include the modernized teacher development framew ork w ith the core ideas of optimizing the teachers’perception,teachers’information literacy,and diversified teaching methods,in w hich the concepts of teachers’self-development and life-long learning are emphasized.M oreover,the four important supporting contents w ere proposed:the teacher autonomy,the construction of intelligent teacher development platform,perfecting the training organization,and the informationization of teachers’training.展开更多
目前层次型或深度模糊系统性能优异,但是模型复杂度较高;而基于蒸馏学习的轻量型TSK(Takagi-Sugeno-Kang)模糊分类器主要以单教师知识蒸馏为主,若教师模型表现不佳,则会影响蒸馏效果和模型的整体性能;此外,传统的多教师蒸馏通常使用无...目前层次型或深度模糊系统性能优异,但是模型复杂度较高;而基于蒸馏学习的轻量型TSK(Takagi-Sugeno-Kang)模糊分类器主要以单教师知识蒸馏为主,若教师模型表现不佳,则会影响蒸馏效果和模型的整体性能;此外,传统的多教师蒸馏通常使用无标签策略分配教师模型输出的权重,容易使低质量教师误导学生。对此,本文提出了一种基于多教师自适应知识蒸馏的TSK模糊分类器(TSK fuzzy classifier based on multi-teacher adaptive knowledge distillation,TSK-MTAKD),以多个具有不同神经表达能力的深度神经网络为教师模型,利用本文提出的多教师知识蒸馏框架从多个深度学习模型中提取隐藏知识,并传递给具有强大不确定处理能力的TSK模糊系统。同时设计自适应权重分配器,将教师模型的输出与真实标签做交叉熵处理,更接近真实值的输出将被赋予更高权重,提高了模型的鲁棒性与隐藏知识的有效性。在13个UCI数据集上的实验结果充分验证了TSK-MTAKD的优势。展开更多
文摘The advantages and disadvantages for learning English in the Network-based environment attract most researchers’concern nowadays.This study profiles college English teachers’beliefs about the networkbased language learning.The main finding is that teachers’beliefs about network-based language learning are heterogeneous and thus reflect a wide range in terms of the evolution of approaches and technology use.
文摘人体行为识别(Human Activity Recognition,HAR)是当前众多研究工作的基石,对于推动人机交互和智能数字化转型具有巨大潜力。由于目标域样本较难采集,现有方法在跨域识别方面表现不佳。为解决这一问题,提出一种新的WiFi使能跨域HAR方法,从WiFi信号中获取信道状态信息(Channel State Information,CSI)并转化为图像,在基于Wasserstein距离和梯度的生成对抗网络(Wasserstein Generative Adversarial Network with Gradient Penalty,WGAN-GP)中引入双判别器,通过与源域样本和单目标域样本特征联合对抗,生成同时带有双域特征的虚拟样本。该方法还结合基于Mean Teacher的半监督学习设计识别分类(Recognition and Classification,RC)模块,通过对有标记样本与无标记样本分别构造损失函数,进行整体一致性损失的评估,实现对目标域样本的识别。实验结果证明了所提方法能够在减轻目标域样本采集压力的同时,实现较高的检测精度,在手势与动作的数据集上测试准确率分别达到92.71%和86.65%。
基金funded by English Teaching Reform in Colleges and Universities Project of Heilongjiang Province titled with The Research on the Application of College English Teaching Mode to Local Colleges and Universities in the Background of MOOC."(0115004)The Key Project of Education and Teaching Research of Jiamusi University(JYZW2015-07)+1 种基金Teaching and Research Project of Jiamusi University(JYWG2014-02)The Interdisciplinary Scientific Research Project(12J201508)
文摘Based on the previous research results and five principles,this paper tries to build an applicable optimization framew ork of college English classroom teaching,w hich centers on teacher development.According to this framew ork,teacher development should include the modernized teacher development framew ork w ith the core ideas of optimizing the teachers’perception,teachers’information literacy,and diversified teaching methods,in w hich the concepts of teachers’self-development and life-long learning are emphasized.M oreover,the four important supporting contents w ere proposed:the teacher autonomy,the construction of intelligent teacher development platform,perfecting the training organization,and the informationization of teachers’training.
文摘目前层次型或深度模糊系统性能优异,但是模型复杂度较高;而基于蒸馏学习的轻量型TSK(Takagi-Sugeno-Kang)模糊分类器主要以单教师知识蒸馏为主,若教师模型表现不佳,则会影响蒸馏效果和模型的整体性能;此外,传统的多教师蒸馏通常使用无标签策略分配教师模型输出的权重,容易使低质量教师误导学生。对此,本文提出了一种基于多教师自适应知识蒸馏的TSK模糊分类器(TSK fuzzy classifier based on multi-teacher adaptive knowledge distillation,TSK-MTAKD),以多个具有不同神经表达能力的深度神经网络为教师模型,利用本文提出的多教师知识蒸馏框架从多个深度学习模型中提取隐藏知识,并传递给具有强大不确定处理能力的TSK模糊系统。同时设计自适应权重分配器,将教师模型的输出与真实标签做交叉熵处理,更接近真实值的输出将被赋予更高权重,提高了模型的鲁棒性与隐藏知识的有效性。在13个UCI数据集上的实验结果充分验证了TSK-MTAKD的优势。
