[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fO7vzt5WepIjfckrBjSjvD8IAy8wEK-BsVEtZ_objmWw":3},{"answer":4,"createTime":5,"id":6,"options":7,"origin":12,"question":19,"related":20,"source":30,"type":31},[],"2024-06-19 10:41:55",153136252,[8,9,10,11],"1","2","3","n",{"count":13,"courseId":14,"courseImg":15,"courseName":16,"workId":17,"workName":18},18,"e435984f691240c6fc7eed81e23d857c","https:\u002F\u002Ftihai-oss-cloud.itihey.com\u002Fimg\u002F9c1e48361b00f3ee2086f4e259ed792b.jpg","语音识别综合实践案例","work_35262542","第五次作业","n-gram模型表示()词之间的组合概率",[21,32,41,50,59,68,76,85,94,97],{"answer":22,"createTime":5,"id":23,"options":24,"question":29,"source":30,"type":31},[],153136244,[25,26,27,28],"语音识别","语音合成","语音增强","说话人识别","在语音领域,均方误差损失函数主要用于()","v1",0,{"answer":33,"createTime":5,"id":34,"options":35,"question":40,"source":30,"type":31},[],153136245,[36,37,38,39],"梯度裁剪","权重衰减","参数归一化","随机丢弃","L1和L2正则化方法也叫做()",{"answer":42,"createTime":5,"id":43,"options":44,"question":49,"source":30,"type":31},[],153136246,[45,46,47,48],"1:1","n:m","1:n","n:1","DNN-HMM的输出节点与HMM的状态节点是()关系",{"answer":51,"createTime":5,"id":52,"options":53,"question":58,"source":30,"type":31},[],153136247,[54,55,56,57],"351","390","350","389","DNN的优势在于能够对包含多帧数据的特征进行处理,表征更丰富的语音上下文特征,DNN的输入层往往先进行多帧拼接,加入每帧语音特征39维度,对前10帧进行拼接,拼接后的特征有几维()",{"answer":60,"createTime":5,"id":61,"options":62,"question":67,"source":30,"type":31},[],153136248,[63,64,65,66],"MFCC","PLP","FBANK","以上都不对","以下哪种声学特征适合基于深度学习的语音模型()",{"answer":69,"createTime":5,"id":70,"options":71,"question":75,"source":30,"type":31},[],153136249,[72,73,74,66],"人工打标","基于其它模型输出的帧与状态的对齐文件","基于人工输出的帧与状态的对齐文件","DNN-HMM语音识别模型是有监督学习,真实标签是从哪里获取到的()",{"answer":77,"createTime":5,"id":78,"options":79,"question":84,"source":30,"type":31},[],153136250,[80,81,82,83],"1,250","1,150","50,250","50,150","假如有50个音素,每个音素有使用5个状态建模,每个音素有效发射状态是3个,那么要训练()个GMM-HMM模型,DNN-HMM的输出节点有多少个()",{"answer":86,"createTime":5,"id":87,"options":88,"question":93,"source":30,"type":31},[],153136251,[89,90,91,92],"(1,0,1)","(-1,0,-1)","(-1,0,1)","(-1,1,0)","TDNN描述当前帧与过去1帧、未来1帧拼接为3帧的节点作为输入节点的表示是()",{"answer":95,"createTime":5,"id":6,"options":96,"question":19,"source":30,"type":31},[],[8,9,10,11],{"answer":98,"createTime":99,"id":100,"options":101,"question":106,"source":30,"type":31},[],"2024-06-19 10:41:56",153136253,[102,103,104,105],"ngram-count -read sample.count -order 3 -lm sample.lm","ngram-count -read sample.count -order 3 -lm sample_wb.lm -wbdiscount","ngram-count -read sample.count -order 3 -lm sample_kn.lm -interpolate -kndiscount","ngram-count -text sample.txt -order 3 -write sample.count","使用SPILM训练语言模型时,如果语料太少,可以使用以下哪种方式训练()"]