[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$f4DHccpclJfrnhfSIX_G_f-jPCeIkIR05xIXtcO9V2ms":3},{"answer":4,"createTime":5,"id":6,"options":7,"origin":12,"question":8,"related":19,"source":29,"type":30},[],"2025-04-09 08:48:42",182075758,[8,9,10,11],"下列哪种算法主要用于将无标签数据组织成不同簇","线性回归","逻辑回归","支持向量机",{"count":13,"courseId":14,"courseImg":15,"courseName":16,"workId":17,"workName":18},55,"e2f50b2f7e53141ca8dc251b2cba11b4","https:\u002F\u002Ftihai-oss-cloud.itihey.com\u002Fimg\u002Fa096acd23dd8acdfe38f1bd3183c74c3.png","安徽省人工智能通识课（AI+心理学）","32c7dd5d3c894d6193e4ce3cc759d697","第三章作业",[20,31,39,48,57,60,69,78,87,96],{"answer":21,"createTime":5,"id":22,"options":23,"question":28,"source":29,"type":30},[],182075754,[24,25,26,27],"机器学习是利用大数据进行统计分析的方法","机器学习是用计算机程序模拟人的学习能力,从数据中提取知识和经验","机器学习仅用于图像和语音识别","机器学习只关注模型的优化问题","根据所学知识,以下哪项最准确地描述了机器学习的定义","v1",0,{"answer":32,"createTime":5,"id":33,"options":34,"question":38,"source":29,"type":30},[],182075755,[35,36,37,10],"K-近邻算法","决策树","主成分分析(PCA)","在机器学习中常用于降维的经典算法是",{"answer":40,"createTime":5,"id":41,"options":42,"question":47,"source":29,"type":30},[],182075756,[43,44,45,46],"迁移学习是指从零开始训练一个模型,以适应新任务","迁移学习是利用已有知识和预训练模型,快速适应新任务并减少数据与计算资源需求的方法","迁移学习仅适用于图像分类领域","迁移学习是专门用来解决领域差异问题的一种算法","关于迁移学习的定义,下列哪项描述最为准确",{"answer":49,"createTime":5,"id":50,"options":51,"question":56,"source":29,"type":30},[],182075757,[52,53,54,55],"批量梯度下降(BGD)每次更新参数时只使用一个样本","随机梯度下降(SGD)使用所有训练样本计算梯度","小批量梯度下降(MBGD)在每次更新时使用部分样本,兼具训练速度和全局最优的优势","梯度下降方法不需要计算损失函数","课程中所提到的梯度下降方法中,下列哪项描述正确",{"answer":58,"createTime":5,"id":6,"options":59,"question":8,"source":29,"type":30},[],[8,9,10,11],{"answer":61,"createTime":5,"id":62,"options":63,"question":68,"source":29,"type":30},[],182075759,[64,65,66,67],"迁移学习不需要选择合适的预训练模型,因为任何模型都可以直接用于新任务","领域差异过大可能导致预训练模型在目标任务上表现不佳,是迁移学习的一大挑战","迁移学习的微调过程完全自动化,不需要人为调整","在迁移学习中,数据需求总是充足的,因此无需担心数据稀缺问题","关于迁移学习面临的挑战,下列哪项描述是正确的",{"answer":70,"createTime":5,"id":71,"options":72,"question":77,"source":29,"type":30},[],182075760,[73,74,75,76],"提高模型在新任务上的收敛速度","提高模型在原任务上的准确率","提高模型的泛化能力","减少模型的参数数量","迁移学习的主要目的是什么",{"answer":79,"createTime":5,"id":80,"options":81,"question":86,"source":29,"type":30},[],182075761,[82,83,84,85],"小数据集","大数据集","无标注数据集","与目标任务相同的数据集","在迁移学习中,预训练模型通常在什么数据集上进行训练",{"answer":88,"createTime":5,"id":89,"options":90,"question":95,"source":29,"type":30},[],182075762,[91,92,93,94],"在新任务上重新训练整个模型","只在新任务上训练模型的最后一层","调整模型的超参数","对预训练模型的部分或全部参数进行微调","迁移学习中的"微调(Fine-tuning)"指的是什么",{"answer":97,"createTime":5,"id":98,"options":99,"question":104,"source":29,"type":30},[],182075763,[100,101,102,103],"A.特征提取只训练模型的最后几层,微调训练整个模型","特征提取训练整个模型,微调只训练最后几层","特征提取和微调没有区别","特征提取不更新预训练模型的权重,微调更新权重","迁移学习中,特征提取(Feature Extraction)和微调(Fine-tuning)的主要区别是什么"]