[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fzad-_F-W59zSjWVH_xODSnlfHIbEWJT3wJBZd_hUkKg":3},{"answer":4,"createTime":5,"id":6,"options":7,"origin":8,"question":15,"related":16,"source":27,"type":95},[],"2025-11-22 22:57:39",239522519,[],{"count":9,"courseId":10,"courseImg":11,"courseName":12,"workId":13,"workName":14},21,"53e1d2ef4961cca8eea3e23969ad2cb9","https:\u002F\u002Ftihai-oss-cloud.itihey.com\u002Fimg\u002F03a579384a6dc297c89809b582fcc767.png","默认课程","work_48082556","第九次课后练习","LSTM网络的三大门控分别是 ____门、 ____门和 ____门",[17,29,38,47,56,65,74,83,92,96],{"answer":18,"createTime":19,"id":20,"options":21,"question":26,"source":27,"type":28},[],"2025-11-22 22:55:03",239521602,[22,23,24,25],"存储当前输入","存储序列的历史信息","直接输出预测结果","计算梯度值","RNN的核心结构中,隐藏状态(Hidden State)的主要作用是( )","v1",0,{"answer":30,"createTime":19,"id":31,"options":32,"question":37,"source":27,"type":28},[],239521603,[33,34,35,36],"序列长度过短","激活函数线性化","长期依赖导致梯度在反向传播中指数衰减","参数矩阵维度过大","RNN在训练中容易出现的梯度消失问题的主要原因是( )",{"answer":39,"createTime":19,"id":40,"options":41,"question":46,"source":27,"type":28},[],239521604,[42,43,44,45],"单一激活函数","门控机制","固定大小输入","卷积层","LSTM网络为了解决RNN的梯度消失问题,引入了( )",{"answer":48,"createTime":19,"id":49,"options":50,"question":55,"source":27,"type":28},[],239521605,[51,52,53,54],"决定新信息的添加","控制输出状态","决定遗忘先前细胞状态的信息","直接生成隐藏状态","在LSTM的三大门控机制中,遗忘门的主要作用是( )",{"answer":57,"createTime":19,"id":58,"options":59,"question":64,"source":27,"type":28},[],239521606,[60,61,62,63],"固定输出大小","共享参数处理变长序列","忽略时间依赖","仅处理图像数据","RNN处理序列数据时,使用循环结构的主要优势是( )",{"answer":66,"createTime":19,"id":67,"options":68,"question":73,"source":27,"type":28},[],239521607,[69,70,71,72],"产生高维稀疏向量,便于词频统计","直接生成稠密嵌入","忽略词汇表大小","加速梯度爆炸","文本数据预处理中,向量化表示常用one-hot编码的原因是( )",{"answer":75,"createTime":19,"id":76,"options":77,"question":82,"source":27,"type":28},[],239521608,[78,79,80,81],"完全相同","LSTM参数是SimpleRNN的4倍","SimpleRNN参数更多","不涉及权重矩阵","LSTM与SimpleRNN在参数数量上的区别是( )",{"answer":84,"createTime":19,"id":85,"options":86,"question":91,"source":27,"type":28},[],239521609,[87,88,89,90],"机器翻译","情感分析","固定尺寸图像分类","文本生成","RNN的应用领域不包括( )",{"answer":93,"createTime":5,"id":6,"options":94,"question":15,"source":27,"type":95},[],[],2,{"answer":97,"createTime":5,"id":98,"options":99,"question":100,"source":27,"type":95},[],239522520,[],"文本预处理中,序列对齐常用 ____技术处理变长序列,如填充至固定长度"]