[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$f0-1BmrKfhuLfJefYuuadqHQLvHgc3U361Vv17kpllbQ":3},{"answer":4,"createTime":5,"id":6,"options":7,"origin":12,"question":16,"related":17,"source":28,"type":29},[],"2026-01-06 23:12:24",284714495,[8,9,10,11],"预设的固定规则","人工定义的特征提取器","从数据中自动学习表征","符号逻辑推理",{"courseId":13,"courseImg":14,"courseName":15},"53e1d2ef4961cca8eea3e23969ad2cb9","https:\u002F\u002Ftihai-oss-cloud.itihey.com\u002Fimg\u002F03a579384a6dc297c89809b582fcc767.png","默认课程","深度学习是人工智能的一个分支,它主要通过什么来学习和改进模型性能",[18,30,39,42,50,59,68,77,86,95],{"answer":19,"createTime":20,"id":21,"options":22,"question":27,"source":28,"type":29},[],"2025-12-24 21:22:58",284714493,[23,24,25,26],"模型存储不需要占用存储空间","模型部署到设备上后不消耗能源","模型不需要电力即可运行","训练大型深度学习模型需要消耗大量电力","深度学习面临的 "能源消耗高" 挑战,主要体现在( )","v1",0,{"answer":31,"createTime":20,"id":32,"options":33,"question":38,"source":28,"type":29},[],284714494,[34,35,36,37],"生成对抗网络","自编码器进行数据压缩","循环神经网络处理时间序列数据","卷积神经网络处理图像数据","校园健康监测APP使用传感器数据预测学生压力水平,这种任务最适合使用",{"answer":40,"createTime":5,"id":6,"options":41,"question":16,"source":28,"type":29},[],[8,9,10,11],{"answer":43,"createTime":5,"id":44,"options":45,"question":49,"source":28,"type":29},[],284714496,[46,47,34,48],"RNN","CNN","自注意力机制",""AI 梵高"网站把简笔画变成油画,关键算法是( )",{"answer":51,"createTime":20,"id":52,"options":53,"question":58,"source":28,"type":29},[],284714497,[54,55,56,57],"循环神经网络(RNN)","卷积神经网络(CNN)","径向基函数网络(RBF)","多层感知机(MLP)","哪种神经网络结构最适合处理像句子这样的序列数据,如分析学生评论文本的情感",{"answer":60,"createTime":5,"id":61,"options":62,"question":67,"source":28,"type":29},[],284714498,[63,64,65,66],"删除模型的无用信息","接收模型的输入数据","存储过去时刻的信息,供当前时刻计算使用","直接输出模型的预测结果","循环神经网络(RNN)的 "隐藏状态" 主要作用是( )",{"answer":69,"createTime":5,"id":70,"options":71,"question":76,"source":28,"type":29},[],284714499,[72,73,74,75],"只能提取一种类型的特征","减少模型的计算量","提取多种不同类型的特征(如边缘、纹理、颜色)","降低模型的泛化能力","卷积神经网络(CNN)中,"多通道卷积" 的作用是( )",{"answer":78,"createTime":20,"id":79,"options":80,"question":85,"source":28,"type":29},[],284714500,[81,82,83,84],"图像风格迁移","智能问答系统","机器翻译","情感分析(判断文本情绪)","深度学习在自然语言处理领域的应用不包括( )",{"answer":87,"createTime":20,"id":88,"options":89,"question":94,"source":28,"type":29},[],284714501,[90,91,92,93],"初始化模型参数","可视化网络结构","计算损失函数对各参数的梯度","前向计算预测结果","在神经网络中,反向传播算法的主要作用是",{"answer":96,"createTime":20,"id":97,"options":98,"question":103,"source":28,"type":29},[],284714502,[99,100,101,102],"数据标注需要大量人力和时间,成本较高","高质量的标注数据能提升模型性能","所有深度学习任务都不需要数据标注","数据标注是给原始数据添加标签(如给图片标注 "猫" 或 "狗")","下列关于深度学习 "数据标注" 的说法,错误的是( )"]