[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fkZxyj0QTOKaAHtdBpGAmUOs1_WRlwUZzBewAP3gwpF0":3},{"answer":4,"createTime":5,"id":6,"options":7,"origin":8,"question":12,"related":13,"source":23,"type":104},[],"2026-01-06 16:56:36",304679577,[],{"courseId":9,"courseImg":10,"courseName":11},"53e1d2ef4961cca8eea3e23969ad2cb9","https:\u002F\u002Ftihai-oss-cloud.itihey.com\u002Fimg\u002F03a579384a6dc297c89809b582fcc767.png","默认课程","生成对抗网络在____(如 AI 绘画)和数据增强(如数据增强)等场景中展现出独特优势",[14,25,34,43,52,61,70,77,86,95],{"answer":15,"createTime":5,"id":16,"options":17,"question":22,"source":23,"type":24},[],304679545,[18,19,20,21],"1943 年 M-P 模型提出","2012 年 AlexNet 在 ImageNet 挑战赛夺冠","1986 年 BP 算法诞生","2006 年深度信念网络提出","标志着深度学习正式登上历史舞台的事件是?( )","v1",0,{"answer":26,"createTime":5,"id":27,"options":28,"question":33,"source":23,"type":24},[],304679546,[29,30,31,32],"无需人工设计特征,实现端到端自动特征学习","依赖浅层模型架构处理复杂数据","通过深层网络结构突破维度灾难瓶颈","在图像识别等任务中达到超越人类的性能","与传统机器学习相比,深度学习的核心优势不包括?( )",{"answer":35,"createTime":5,"id":36,"options":37,"question":42,"source":23,"type":24},[],304679547,[38,39,40,41],"无法处理线性分类问题","缺乏自主学习功能","只能处理图像数据","参数量过大","M-P 模型作为深度学习的早期基础,其主要缺陷是?( )",{"answer":44,"createTime":5,"id":45,"options":46,"question":51,"source":23,"type":24},[],304679548,[47,48,49,50],"完全基于注意力机制,摒弃循环和卷积结构","采用多层感知器实现特征提取","引入池化层降低维度","依赖时序循环连接建模","Transformer 架构的核心创新是?( )",{"answer":53,"createTime":5,"id":54,"options":55,"question":60,"source":23,"type":24},[],304679549,[56,57,58,59],"无法处理图像数据","虚假信息生成与传播","参数量过小","只能运行在本地设备","生成式人工智能面临的首要风险是?( )",{"answer":62,"createTime":5,"id":63,"options":64,"question":69,"source":23,"type":24},[],304679550,[65,66,67,68],"记忆遗忘","选择性关注","逻辑推理","语言生成","注意力机制的工作原理类似于人类的?( )",{"answer":71,"createTime":5,"id":72,"options":73,"question":76,"source":23,"type":24},[],304679551,[21,20,74,75],"1998 年 LeNet-5 提出","2012 年 AlexNet 夺冠","深度学习第三次热潮的起点是?( )",{"answer":78,"createTime":5,"id":79,"options":80,"question":85,"source":23,"type":24},[],304679552,[81,82,83,84],"图像分类","时序数据处理(语音、文本)","特征降维","强化学习","循环神经网络(RNN)的核心应用场景是?( )",{"answer":87,"createTime":5,"id":88,"options":89,"question":94,"source":23,"type":24},[],304679553,[90,91,92,93],"技术迭代与市场需求的矛盾","发展创新与风险管控的平衡","硬件成本与算法精度的矛盾","数据规模与存储能力的矛盾","人工智能伦理治理的核心矛盾是?( )",{"answer":96,"createTime":5,"id":97,"options":98,"question":103,"source":23,"type":24},[],304679554,[99,100,101,102],"仅深度学习需要数据训练","深度学习通过深层网络自动提取特征","机器学习无法处理复杂任务","深度学习不依赖算法","深度学习与机器学习的本质区别在于?( )",2]