[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fO9XEqBe-L99WjbWnWAP7lAG3ezhDSmrrD6ErfuVdrKo":3},{"answer":4,"createTime":5,"id":6,"options":7,"origin":10,"question":16,"related":17,"source":21,"type":22},[],"2023-05-07 20:40:40",33447912,[8,9],"对","错",{"courseId":11,"courseImg":12,"courseName":13,"workId":14,"workName":15},"1000009004","https:\u002F\u002Ftihai-oss-cloud.itihey.com\u002Fimg\u002Ff4e4dd841924a91dc656054650011ffd.jpeg","人工智能基础导学","14774745","第七章单元测试","v对于一个样本,如果当前权重能够正确判断其类型,就减小当前权重",[18,23,28,33,38,48,53,58,63,72],{"answer":19,"createTime":5,"id":6,"options":20,"question":16,"source":21,"type":22},[],[8,9],"v2",3,{"answer":24,"createTime":5,"id":25,"options":26,"question":27,"source":21,"type":22},[],33447922,[8,9],"神经网络是由一个神经元构成的",{"answer":29,"createTime":5,"id":30,"options":31,"question":32,"source":21,"type":22},[],33447940,[8,9],"隐含层,是指其中神经元的状态在输出端无法直接观测",{"answer":34,"createTime":5,"id":35,"options":36,"question":37,"source":21,"type":22},[],33448035,[8,9],"用生成式模型根据少量样本来估计整个类型的概率特征是很困难的",{"answer":39,"createTime":5,"id":40,"options":41,"question":46,"source":21,"type":47},[],33448046,[42,43,44,45],"二者都是由多个神经元组成的多层神经网络","输入信号向后传递的过程中,都是加权和的计算","二者学习的关键都是神经元的损失计算","二者都有输入、激活和输出","以下说法中,不属于感知器和FNN模型的相同点的是",0,{"answer":49,"createTime":5,"id":50,"options":51,"question":52,"source":21,"type":22},[],33448069,[8,9],"生成式模型模拟概率分布时,常用"后验分布"",{"answer":54,"createTime":5,"id":55,"options":56,"question":57,"source":21,"type":22},[],33448073,[8,9],"生成对抗网络结合了生成模型和判别模型",{"answer":59,"createTime":5,"id":60,"options":61,"question":62,"source":21,"type":22},[],33448103,[8,9],"判别式模型对问题本质缺乏了解,无法从个例中抽象出整体概念",{"answer":64,"createTime":5,"id":65,"options":66,"question":71,"source":21,"type":47},[],33448132,[67,68,69,70],"FNN的输出结果只能是向量","在FNN中,输入信号的传递方向是明确的,并不存在反向信号传递","一个标准的前馈神经网络只有一个输入层和一个输出层","FNN的同层神经元之间存在连接","以下关于前馈神经网络(FNN)的说法正确的是",{"answer":73,"createTime":5,"id":74,"options":75,"question":80,"source":21,"type":47},[],33448215,[76,77,78,79],"重调整采用"奖惩分明"策略,即对于能够准确判断样本类型的权重,提高当前权重比例,反之则降低当前权重比例","感知器模型的关键,就是通过调整权重使一类样本可以激活神经元,而另一类则不会","一层感知器只能针对线性可分的数据集分类,无法解决异或(XOR)问题","感知器模型中的激活函数是二值函数时,损失函数是可导的","以下关于感知器的说法错误的是"]